Association Study Of Dysfunctional Attitudes And Executive Function With Depressive Symptoms Among High School Students

Objectives:(1)To examine the longitudinal associations of dysfunctional attitudes(DA)with the risk of depressive symptoms among high school students.(2)To explore an interaction between DA and executive function on depressive symptoms.MethodsFirst,a cross-sectional study was conducted.A total of 3941 students were recruited from four high schools in Hefei by stratified-cluster sampling method.The survey was assessed on depressive symptoms and DA in April 2015,to preliminary analysis the relationship between DA and depressive symptoms,and their relevant influencing factors in order to provide the basis for the potential causal relationship between DA and depressive symptoms in next study.Secondly,a prospective cohort study was implemented based on the results of cross-sectional study.All high school students in grade one from five schools in Hefei were recruited by stratified-cluster sampling method,and 2493 students without depressive symptoms were screened for follow up subjects by using the Center for Epidemiological Studies Depression Scale(CES-D).Two years prospective cohort study was been carried out to determine effects of DA on the risk of depressive symptoms(follow-up 3 times,6 months for each follow up interval,the first follow-up in April 2016).Finally,2067 valid samples though all of three follow up were included in the analysis sample size.Finally,nested case-control design within a cohort of DA and depressive symptoms was been applied.75 students with new-onset depressive symptoms screened during the first and second follow-up were selected as the case group,and the control group were matched to numbers of students,gender,and class.Attention Networks Test(ANT),Wisconsin Card Sorting Test(WCST),Tower of London(TOL),Digit Span(DS)were conducted among 150 students with and without new-onset depressive symptoms.The main survey tools used in this study:depressive symptoms were assessed via the Center for Epidemiological Studies Depression Scale(CES-D),and depressive symptoms were defined as CES-D?24.Dysfunctional attitudes were assessed via the Dysfunctional Attitudes Scales(DAS),and scores of DA were divided into quartiles(Q1,Q2,Q3,and Q4)to analyze the relationship between it and depressive symptoms.The main statistical analysis methods included that longitudinal correlation analysis between depressive symptoms and DA using generalized linear mixed model(GLMM)and receiver operating characteristic curve(ROC),changes of DA and each dimension in four measurements analysis by the repeated measurement ANOVA and the internal consistency test with Cronbach's alpha,Combined effect of executive function and DA on risk of depressive symptoms via the multiplicative and additive interaction analysis.ResultsCross-sectional survey from 3941 high school students showed that the prevalence of depressive symptoms was 33.1%.When scores of DA were divided into quartiles(Q1,Q2,Q3,and Q4),the prevalences of depressive symptoms were 12.4%,23.8%,34.3%,and 61.9%in Q1?Q2?Q3?Q4 group,respectively,and there was a significantly linear trend(?~2_{for trend}=566.199,P<0.001).Multivariate logistic regression analysis showed that DA was closely related to depressive symptoms(P(27)0.05),compared with Q1 of DA,the top three quartiles(Q2,Q3 and Q4)exhibited increased risk for depressive symptoms(Q2 OR=1.91;Q3 OR=2.97;Q4 OR=7.95)(all P(27)0.01).Furthermore,four dimensions of DA(D1:vulnerability;D2:attraction and repulsion;D3:perfectionism;D8:cognitivephilosophy)were also significantly correlated with depressive symptoms.Family function,life events,and self-esteem were associated factors with DA and depressive symptoms.Two years prospective cohort study from 2067 students without depressive symptoms found that(1)the incidences of depressive symptoms in the first,second and third follow-up were 28.2%,23.8%,and 18.7%,respectively,and 7.3%of students reported depressive symptoms through three follow-up visits.(2)The longitudinal associations of DA with the risk of depressive symptoms using GLMM analysis.Compared with Q1of DA,GLMM showed that Q2(RR=1.73,95%CI:1.38?2.17),Q3(RR=2.52,95%CI:2.02?3.15)and Q4(RR=4.43,95%CI:3.57?5.51)groups had elevated risk of depressive symptoms,and there was a significant linear correlation between DA and depressive symptoms(all P(27)0.001).(3)The longitudinal associations between DA dimensions and depressive symptoms using GLMM analysis:There were significant correlations between depressive symptoms and five DA dimensions(D1:vulnerability;D2:attraction and repulsion;D3:perfectionism;D5:seeking applause;D8:cognitivephilosophy),and D3 was associated with depressive symptoms only in girls.(4)The predicting role of baseline DA score for depressive symptoms using ROC analysis.AUC of ROC in the first,second and third follow-up were 0.587(95%CI:0.559?0.614),0.566(95%CI:0.537?0.594)and 0.595(95%CI:0.564?0.626),respectively,and the AUC of ROC among students with depressive symptoms occurring in all of three follow-ups was 0.609(95%CI:0.564?0.654).Furthermore,AUC of baseline DA,D1(vulnerability),and D2(attraction and repulsion)scores predicting for depressive symptoms occurring in all of three follow-ups were 0.804(95%CI:0.690?0.919),0.705(95%CI:0.529?0.880)and 0.925(95%CI:0.856?0.994)among the students with higher levels of life events(above P₅₀ of scores),respectively.But those were not found in the students with lower levels of life events(below P₅₀ of scores).(5)Changes of DA and each dimension in four measurements.Repeated measurement ANOVA showed that there were significant differences of DA score in four measurements(P<0.001).The scores of DA,D1,D2,and D8 in each follow-up were significantly higher than those in baseline(P<0.05),while the scores of other five DA dimensions only in the first follow-up was higher than those in baseline(P<0.05).Furthermore,the internal consistency test of the four evaluations of DA(Quartile of DA score)showed that the DA Cronbach's?was 0.806,except that D8 was 0.69,Cronbach's of other DA dimensions were above 0.7.The associations of executive function and DA with the risk of depressive symptoms in high school students using the nested case-control design:(1)Comparisons of neuropsychological characteristics between the case and control group.In DS test,the Reaction Time of the case group(9.38±3.22)was significantly higher than that of the control group(8.41±2.10)(P<0.05).In ANT test,the Executive Control Network in case group[82.08(68.28,99.28)]was significantly higher than that in control group[70.86(46.24,92.72)](P<0.01).In WCST test,the Number of Continuous Answers(51.19±13.77 vs 35.76±11.37)and the Number of Continuous Errors(35.76±11.37 vs24.35±11.02)in case group were significantly higher than those in control group(all P<0.05),and the Number of Classification Completed in case group(3.71±1.90)was lower than that in control group(4.35±1.98)(P<0.05).In TOL test,the Planned Time(5.13±1.41 vs 4.61±1.16),Completion Time(15.32±4.03 vs 14.09±3.01)and Completion Steps(7.52±0.75 vs 7.02±0.74)in case group were significantly higher than those in control group(P<0.05).(2)Associations of executive function and DA with depressive symptoms.Multivariate logistic regression analyses showed that higher DA scores,more Persistent Errors,and lower efficiency of Executive Control Network were risk factors of depressive symptoms.(3)Interaction between executive function and DA on depressive symptoms.There was a multiplicative interaction between Persistent Errors Number and DA on depressive symptoms(P<0.05).The students with higher scores of Persistent Errors Numbe(above P₅₀)and DA(above P₅₀)exhibited significant increased risk for depressive symptoms than those in the students with lower scores of Number of Persistent Errors(below P₅₀)and DA(below P₅₀)(OR=8.70,95%CI:1.19?63.75).But the additive interaction between Persistent Error Number and DA on depressive symptoms was not found(RERI=1.90,95%CI:-12.24?16.03;AP=0.22,95%CI:-1.19?1.62;S=1.33,95%CI:0.17?10.59).Meanwhile,there was a multiplicative interaction between Executive Control Network and DA on depressive symptoms(P<0.05).The students with higher scores of Executive Control Network(above P₅₀)and DA(above P₅₀)exhibited significant increased risk for depressive symptoms than those in the students with lower scores of Executive Control Network(below P₅₀)and DA(below P₅₀)(OR=10.82,95%CI:1.34?87.35).But the additive interaction of Executive Control Network and DA on depressive symptoms was not found(RERI=4.16,95%CI:-13.61?21.94);AP=0.38,95%CI:-0.79?1.56);S=1.74,95%CI:0.19?15.52).Conclusions:Depressive symptoms was common in high school students.We found that the risk of depressive symptoms increases with higher DA level.Among the students with high level of negative life event stress,DA could better predict the recurrent depressive symptoms during 2-year follow-up.The five DA dimensions were closely related to the occurrence of depressive symptoms,and gender difference was found in those associations.Persistent Errors Number and DA,Executive Control Network and DA have a multiplicative interaction on the occurrence of depressive symptoms in high school students.We postulate there was an interaction between DA and executive function on depressive symptoms.