Logistic回归与决策树模型在青少年抑郁焦虑症状共病影响因素中的应用

叶盛; 杨月; 卢雪蕾; 刘辉; 李俊同; 刘黎

doi:10.16462/j.cnki.zhjbkz.2024.08.007

Logistic回归与决策树模型在青少年抑郁焦虑症状共病影响因素中的应用

doi: 10.16462/j.cnki.zhjbkz.2024.08.007

南京市疾病预防控制中心学校卫生科，南京 210003

基金项目:

南京市疾病预防控制中心青年人才科研培育团队项目 NPY2304

详细信息

通讯作者:
刘黎，E-mail : 895047907@qq.com

中图分类号: R181.3
计量
- 文章访问数: 53
- HTML全文浏览量: 23
- PDF下载量: 15
- 被引次数: 0
出版历程
- 收稿日期: 2024-01-30
- 修回日期: 2024-06-16
- 网络出版日期: 2024-09-29
- 刊出日期: 2024-08-10

Application of logistic regression and decision tree model on influencing factors of the comorbidity of depression and anxiety symptoms in adolescents

Section of School Health, Nanjing Municipal Center for Disease Control and Prevention, Nanjing 210003, China

Funds:

Youth Foundation of Nanjing Municipal Center for Disease Control and Prevention NPY2304

More Information

Corresponding author: LIU Li, E-mail : 895047907@qq.com

摘要

摘要: 目的应用logistic回归与决策树模型分析青少年抑郁焦虑症状共病的影响因素，为相关干预措施提供依据。方法 2022年10月，基于立意抽样和分层随机整群抽样法选取南京市8个监测区10 178名青少年作为研究对象，问卷调查青少年抑郁焦虑症状共病及其影响因素。构建logistic回归与决策树模型，应用受试者工作特征(receive operating characteristic, ROC)曲线下面积(area under curve, AUC)评价模型预测效果。结果南京市青少年抑郁焦虑症状共病检出率为16.63%。Logistic回归与决策树模型分析结果均显示，自伤行为、学业满意程度、校园欺凌、学段和饮酒史是青少年抑郁焦虑症状共病的重要影响因素(均P < 0.05)。Logistic回归模型AUC大于决策树模型(0.828∶0.787, Z=11.153, P < 0.01)。结论 Logistic回归模型的预测效果优于决策树模型。应重点关注自伤行为、学业满意程度、校园欺凌对青少年抑郁焦虑症状共病的影响，帮助青少年养成良好的健康行为方式，促进青少年身心健康发展。
- 青少年 /
- 抑郁 /
- 焦虑 /
- 共病 /
- 决策树模型
Abstract: Objective To analyze the influencing factors of the comorbidity of depression and anxiety symptoms among adolescents by using logistic regression and decision tree model, and to provide basis for relevant intervention. Methods In October 2022, a combination of purposive sampling and stratified random cluster sampling method was employed to recruit 10 178 adolescents across eight districts in Nanjing. Data were collected through a questionnaire assessing the comorbidity of depression and anxiety symptoms and potential influencing factors. Logistic regression and decision tree analyses were conducted to identify key predictors, with model performance evaluated using the receiver operating characteristic (ROC) area under curve (AUC). Results The prevalence of the comorbidity of depression and anxiety symptoms was 16.63% among adolescents. Both analytical models highlighted self-harm behavior, academic satisfaction, school bullying, educational phase and drinking history as significant influencing factors of these comorbid symptoms (all P < 0.05). The area under ROC curve of logistic regression model was larger than that of decision tree model (0.828 ∶ 0.787, Z=11.153, P < 0.01). Conclusions The predictive ability of the logistic regression model is better than that of the decision tree model. More attention should be paid to the influence of self-harm, academic satisfaction and school bullying on the comorbidity of depression and anxiety symptoms. Additionally, efforts should be made to help adolescents develop healthy behaviors, thereby promoting their physical and mental well-being.
- Adolescents /
- Depression /
- Anxiety /
- Comorbidity /
- Decision tree model

HTML全文

图 1 青少年抑郁焦虑症状共病的多因素logistic回归分析

Figure 1. Multivariate logistic regression analysis on the comorbidity of depression and anxiety symptoms in adolescents

下载: 全尺寸图片幻灯片

图 2 青少年抑郁焦虑症状共病影响因素的决策树树形图

Figure 2. Decision tree of the factors influencing comorbidity of depression and anxiety symptoms in adolescents

下载: 全尺寸图片幻灯片

图 3 Logistic回归模型和决策树模型的ROC曲线图

ROC：受试者工作特征。

Figure 3. ROC curves of decision tree model and logistic regression model

ROC: receive operating characteristic.

下载: 全尺寸图片幻灯片

表 1 不同个人特征青少年抑郁症状、焦虑症状及抑郁焦虑症状共病检出率

Table 1. Risk rates for depression symptoms, anxiety symptoms and the comorbidity of depression and anxiety symptoms among adolescents with different personal characteristics

变量 Variable	人数 Total (n=10 178)	抑郁症状 Depression symptoms		焦虑症状 Anxiety symptoms		抑郁焦虑症状共病 The comorbidity of depression and anxiety symptoms
变量 Variable	人数 Total (n=10 178)	检出人数 Number of cases ^① (n=2 109)	χ²值 value	检出人数 Number of cases ^① (n=2 828)	χ²值 value	检出人数 Number of cases ^① (n=1 693)	χ²值 value
性别 Gender
男 Male	5 412	1 011(18.68)	29.292 ^②	1 390(25.68)	25.444 ^②	788(14.56)	35.840 ^②
女 Female	4 766	1098(23.04)		1 438(30.17)		905(18.99)
学段 Phase of studying
小学 Primary school	3 889	469(12.06)	363.375 ^②	732(18.82)	369.747 ^②	359(9.23)	327.903 ^②
初中 Junior high school	3 588	797(22.21)		1 005(28.01)		631(17.59)
高中 Senior high school	2 701	843(31.21)		1 091(40.39)		703(26.03)
地区 District
城区 Urban	5 500	916(16.65)	120.465 ^②	1 294(23.53)	108.135 ^②	726(13.20)	101.756 ^②
郊县 Suburban	4 678	1 193(25.50)		1 534(32.79)		967(20.67)
家庭结构 Family structure
核心家庭 Core family	5 510	1 084(19.67)	6.831 ^②	1 503(27.28)	1.278	872(15.83)	4.617
非核心家庭 Non-core family	4 392	958(21.81)		1 243(28.30)		766(17.44)
父亲文化程度 Education level of father
小学及以下 Elementary school or below	439	134(30.52)	117.162 ^②	158(35.99)	90.357 ^②	111(25.28)	92.828 ^②
初中/中专/高中 Junior high school/technical secondary school/high school	6 064	1 414(23.32)		1 848(30.47)		1 132(18.67)
大学及以上 University or above	3 675	561(15.27)		822(22.37)		450(12.24)
母亲文化程度 Education level of mother
小学及以下 Elementary school or below	843	249(29.54)	137.805 ^②	336(39.86)	146.727 ^②	206(24.44)	120.773 ^②
初中/中专/高中 Junior high school/technical secondary school/high school	5 964	1 371(22.99)		1 778(29.81)		1 105(18.53)
大学及以上 University or above	3 371	489(14.51)		714(21.18)		382(11.33)
家庭经济条件 Family economic status
差 Poor	580	232(40.00)	144.307 ^②	265(45.69)	106.819 ^②	202(34.83)	154.747 ^②
中等 Moderate	7 584	1 507(19.87)		2 070(27.29)		1 207(15.92)
好 Good	1 614	284(17.60)		383(23.73)		214(13.26)
不清楚 Unclear	400	86(21.50)		110(27.50)		70(17.50)
超重肥胖 Overweight and obesity
否 No	6 167	1 276(20.69)	0.009	1 718(27.86)	0.041	1 024(16.60)	0.010
是 Yes	4 011	833(20.77)		1 110(27.67)		669(16.68)
学业满意程度 Degree of academic satisfaction
不满意 Dissatisfaction	2 771	1 044(37.68)	808.094 ^②	1 210(43.67)	648.330 ^②	863(31.14)	689.221 ^②
一般 General satisfaction	3 758	748(19.90)		1 072(28.53)		590(15.70)
满意 Satisfaction	3 649	317(8.69)		546(14.96)		240(6.58)
中高强度活动 Moderate-to-vigorous intensity activity/(d·w^-1)
7	1 486	228(15.34)	30.636 ^②	307(20.66)	44.035 ^②	172(11.57)	32.118 ^②
< 7	8 692	1 881(21.64)		2 521(29.00)		1 521(17.50)
屏幕使用时间 Screen time/(h·d^-1)
≤2	3 737	645(17.26)	43.067 ^②	810(21.68)	109.878 ^②	489(13.09)	53.623 ^②
>2	6 441	1 464(22.73)		2 018(31.33)		1 204(18.69)
睡眠时间 Sleep duration
不足 Inadequate	8 289	1 877(22.64)	100.568 ^②	2 512(30.31)	141.327 ^②	1 525(18.40)	100.214 ^②
充足 Adequate	1 889	232(12.28)		316(16.73)		168(8.89)
遭受校园欺凌 School bullying
否 No	9 790	1 854(18.94)	497.250 ^②	2 556(26.11)	360.009 ^②	1 460(14.91)	548.352 ^②
是 Yes	388	255(65.72)		272(70.10)		233(60.05)
自伤行为 Self-injury
无 No	9 071	1 399(15.42)	1 425.233 ^②	2 081(22.94)	975.368 ^②	1 070(11.80)	1 407.773 ^②
有 Yes	1 107	710(64.14)		747(67.48)		623(56.28)
吸烟史History of smoking
无 No	9 772	1 896(19.40)	259.358 ^②	2 593(26.53)	190.893 ^②	1 506(15.41)	264.034 ^②
有 Yes	406	213(52.46)		235(57.88)		187(46.06)
饮酒史 History of drinking
无 No	8 598	1 399(16.27)	667.635 ^②	2 024(23.54)	497.426 ^②	1 092(12.70)	617.917 ^②
有 Yes	1 580	710(44.94)		804(50.89)		601(38.04)
注：学段中的小学仅为四至六年级；家庭结构变量有部分缺失数据。 ①以人数(占比/%)表示; ②P < 0.05。 Note: The option primary school in phase of studying variable was for grades 4 to 6 only；There were missing data for the family structure variable. ①Number of people (proportion/%); ②P < 0.05.

下载: 导出CSV

参考文献(25)

[1]	王萌, 陶芳标, 伍晓艳. 儿童青少年焦虑抑郁共病研究进展[J]. 中华预防医学杂志, 2022, 56(7): 1011-1016. DOI: 10.3760/cma.j.cn112150-20220325-00283. Wang M, Tao FB, Wu XY. Research progress on the comorbidity of anxiety and depression in children and adolescents[J]. Chin J Prev Med, 2022, 56(7): 1011-1016. DOI: 10.3760/cma.j.cn112150-20220325-00283.
[2]	Leyfer O, Gallo KP, Cooper-Vince C, et al. Patterns and predictors of comorbidity of DSM-IV anxiety disorders in a clinical sample of children and adolescents[J]. J Anxiety Disord, 2013, 27(3): 306-311. DOI: 10.1016/j.janxdis.2013.01.010.
[3]	Losada-Baltar A, Márquez-González M, Jiménez-Gonzalo L, et al. Differences in anxiety, sadness, loneliness and comorbid anxiety and sadness as a function of age and self-perceptions of aging during the lock-out period due to COVID-19[J]. Rev Esp Geriatr Gerontol, 2020, 55(5): 272-278. DOI: 10.1016/j.regg.2020.05.005.
[4]	Cummings CM, Caporino NE, Kendall PC. Comorbidity of anxiety and depression in children and adolescents: 20 years after[J]. Psychol Bull, 2014, 140(3): 816-845. DOI: 10.1037/a0034733.
[5]	Choi KW, Kim YK, Jeon HJ. Comorbid anxiety and depression: clinical and conceptual consideration and transdiagnostic treatment[J]. Adv Exp Med Biol, 2020, 1191: 219-235. DOI: 10.1007/978-981-32-9705-0_14.
[6]	曲晨, 覃玉, 毛涛, 等. 决策树模型与logistic回归在中学生尝试吸烟影响因素中的应用[J]. 中国慢性病预防与控制, 2020, 28(4): 264-269. DOI: 10.16386/j.cjpccd.issn.1004-6194.2020.04.006. Qu C, Qin Y, Mao T, et al. Application of decision tree model and logistic regression on influencing factors for trying-smoking of students in middle schools[J]. Chin J Prev Contr Chron Dis, 2020, 28(4): 264-269. DOI: 10.16386/j.cjpccd.issn.1004-6194.2020.04.006
[7]	疾病预防控制局. 中华人民共和国国家卫生健康委办公厅关于印发2019年全国学生常见病和健康影响因素监测与干预工作方案的通知[EB/OL]. (2019-03-26)[2024-01-24]. http://www.nhc.gov.cn/jkj/s5898bm/201903/1bcbac21e1864377ad 24984fac014c7d.shtml.
[8]	张云婷, 马生霞, 陈畅, 等. 中国儿童青少年身体活动指南[J]. 中国循证儿科杂志, 2017, 12(6): 401-409. DOI: 10.3969/j.issn.1673-5501.2017.06.001. Zhang YT, Ma SX, Cheng C, et al. Physical activity guide for children and adolescents in China[J]. Chin J Evid Based Pediatr, 2017, 12(6): 401-409. DOI: 10.3969/j.issn.1673-5501.2017.06.001.
[9]	中华人民共和国教育部. 中华人民共和国教育部等八部门关于印发《综合防控儿童青少年近视实施方案》的通知[EB/OL]. (2018-08-30)[2024-01-24]. http://www.moe.gov.cn/srcsite/A17/moe_943/s3285/201808/t20180830_346672.html.
[10]	李敏辉, 杨亚明, 肖月, 等. 中小学生抑郁焦虑者亚型和风险特征分析[J]. 中国学校卫生, 2023, 44(12): 1793-1798. DOI: 10.16835/j.cnki.1000-9817.2023.12.008. Li MH, Yang YM, Xiao Y, et al. Analysis of subtypes and risk characteristics of depression and anxiety among primary and secondary school students[J]. Chin J Sch Health, 2023, 44(12): 1793-1798. DOI: 10.16835/j.cnki.1000-9817.2023.12.008.
[11]	王浩, 周怡, 戴品远, 等. 浙江省中学生焦虑抑郁症状共病状况分析[J]. 中华流行病学杂志, 2023, 44(12): 1921-1927. DOI: 10.3760/cma.j.cn112338-20230722-00032. Wang H, Zhou Y, Dai PY, et al. Comorbidity of anxiety symptoms and depression symptoms among middle and high school students in Zhejiang Province[J]. Chin J Epidemiol, 2023, 44(12): 1921-1927. DOI: 10.3760/cma.j.cn112338-20230722-00032.
[12]	Wang M, Mou XY, Li TT, et al. Association between comorbid anxiety and depression and health risk behaviors among Chinese adolescents: cross-sectional questionnaire study[J]. JMIR Public Health Surveill, 2023, 9: e46289. DOI: 10.2196/46289.
[13]	Sampasa-Kanyinga H, Colman I, Goldfield GS, et al. Combinations of physical activity, sedentary time, and sleep duration and their associations with depressive symptoms and other mental health problems in children and adolescents: a systematic review[J]. Int J Behav Nutr Phys Act, 2020, 17(1): 72. DOI: 10.1186/s12966-020-00976-x.
[14]	Sarris J, Thomson R, Hargraves F, et al. Multiple lifestyle factors and depressed mood: a cross-sectional and longitudinal analysis of the UK Biobank (N=84, 860)[J]. BMC Med, 2020, 18(1): 354. DOI: 10.1186/s12916-020-01813-5.
[15]	袁勇, 干敏雷, 尹小俭, 等. 青少年非自杀性自伤行为与抑郁症状的相关性[J]. 中国学校卫生, 2023, 44(5): 659-663. DOI: 10.16835/j.cnki.1000-9817.2023.05.005. Yuan Y, Gan ML, Yin XJ, et al. Correlation analysis between non-suicidal self-injury behavior and depressive symptoms in adolescents[J]. Chin J Sch Health, 2023, 44(5): 659-663. DOI: 10.16835/j.cnki.1000-9817.2023.05.005.
[16]	Hawton K, Saunders KEA, O′Connor RC. Self-harm and suicide in adolescents[J]. Lancet, 2012, 379(9834): 2373-2382. DOI: 10.1016/S0140-6736(12)60322-5.
[17]	Tilton-Weaver L, Marshall SK, Svensson Y. Depressive symptoms and non-suicidal self-injury during adolescence: Latent patterns of short-term stability and change[J]. J Adolesc, 2019, 75: 163-174. DOI: 10.1016/j.adolescence.2019.07.013.
[18]	李金文, 白荣, 王雨萌, 等. 青少年抑郁与自伤行为的发展轨迹及其关系: 基于两年的追踪研究[J]. 心理发展与教育, 2023, 39(3): 429-438. DOI: 10.16187/j.cnki.issn1001-4918.2023.03.14. Li JW, Bai R, Wang YM, et al. The relationship between developmental trajectories of Adolescents′Depression and self-injury: a two-year longitudinal study[J]. Psychol Dev Educ, 2023, 39(3): 429-438. DOI: 10.16187/j.cnki.issn1001-4918.2023.03.14.
[19]	Burke TA, Fox K, Zelkowitz RL, et al. Does nonsuicidal self-injury prospectively predict change in depression and self-criticism?[J]. Cognit Ther Res, 2019, 43(2): 345-353. DOI: 10.1007/s10608-018-9984-8.
[20]	Zhu XH, Haegele JA, Liu HR, et al. Academic stress, physical activity, sleep, and mental health among Chinese adolescents[J]. Int J Environ Res Public Health, 2021, 18(14): 7257. DOI: 10.3390/ijerph18147257.
[21]	Islam MI, Khanam R, Kabir E. Bullying victimization, mental disorders, suicidality and self-harm among Australian high schoolchildren: Evidence from nationwide data[J]. Psychiatry Res, 2020, 292: 113364. DOI: 10.1016/j.psychres.2020.113364.
[22]	Gong ZP, Reinhardt JD, Han ZQ, et al. Associations between school bullying and anxiety in children and adolescents from an ethnic autonomous county in China[J]. Psychiatry Res, 2022, 314: 114649. DOI: 10.1016/j.psychres.2022.114649.
[23]	Moore SE, Norman RE, Suetani S, et al. Consequences of bullying victimization in childhood and adolescence: a systematic review and meta-analysis[J]. World J Psychiatry, 2017, 7(1): 60-76. DOI: 10.5498/wjp.v7.i1.60.
[24]	Skogen JC, Knudsen AK, Hysing M, et al. Trajectories of alcohol use and association with symptoms of depression from early to late adolescence: The Norwegian Longitudinal Health Behaviour Study[J]. Drug Alcohol Rev, 2016, 35(3): 307-316. DOI: 10.1111/dar.12350.
[25]	Hurlocker MC, Carlon H, Pearson MR, et al. Trajectories of change in subclinical anxiety and alcohol use during alcohol treatment: a parallel process growth model[J]. Drug Alcohol Depend, 2023, 246: 109838. DOI: 10.1016/j.drugalcdep.2023.109838.