Interactive effects of delayed bedtime and family-associated factors on depression in elementary school children
Jin-Ding Lin
a, Ho-Jui Tung
b, Yu-Hsin Hsieh
a, Fu-Gong Lin
a,*
aSchoolofPublicHealth,NationalDefenseMedicalCenter,Taipei,Taiwan
bDepartmentofHealthcareAdministration,AsiaUniversity,Taichung,Taiwan
1. Introduction
Sleepisoneofthecriticalfactorstoschoolchildrenfortheirphysiologicalwell-beingandqualityoflife.Lackofsleepmay impacttheirdaytimeactivitiesandacademicperformanceinschool(Dewald,Meijer,Oort,Kerkhof,&Bogels,2010;Hudson, Gradisar,Gamble,Schniering,&Rebelo,2009).Theaveragedailysleepdurationofelementaryschool-agedchildrenhasbeen foundtorangefrom9.3to10.6h(Iglowstein,Jenni,Molinari,&Largo,2003).Insufficientsleephasbeenfoundtobecommon forchildrenandadolescentsinbothAsianandWesterncountries(Lue,Wu,&Yen,2010;Mooreetal.,2009).Sleepproblems inchildrenmayleadtoanxietyanddepression(Hudsonetal.,2009;Mooreetal.,2009).
ARTICLE INFO
Articlehistory:
Received11August2011 Accepted11August2011 Availableonline9September2011
Keywords:
Sleepduration Bedtime Depression Interactioneffect
ABSTRACT
Shortersleep timewasreportedtobeassociatedwith psychologicalfunctioningin children.Weintendedtoexaminetherelationshipbetweennocturnalsleepdurationand depressionstatusbyinvestigatingifdelayedbedtimecouldbeoneoftheenhancement factorsfordepressioninchildren.Across-sectionalstudywasperformedtoinvestigate the depression status in elementary school children in middle Taiwan. Total 676 participantsfrom29schools,ingrades3–6wererecruitedtoparticipateinthestudy.A modified depressionscale fordomestic schoolchildrenwas usedtodeterminethe depression status. Data of depression-related demographic characteristics, family, schoolvariablesandbedtimedatawerecollectedwithastructuredquestionnaire.The resultsshowedthatalmostoneinfivechildren(18%)haddepressionstatus.Delayed bedtime,child–parentrelationship,familyclimate,andpeerrelationshipwerefoundto bethemainpredictorsofchildhooddepression.Furtherstratificationanalysisshowed thatdelayedbedtimesignificantlyinteractedwithfamilyclimateandpeerrelationship onchildhooddepression.Theriskofdepressionforchildrenwithadelayedbedtimeof10 PM and either in a non-harmonious family life or without a close parent–child relationshipwas4.35and4.73timesgreaterthanthereferencegrouprespectively.This studyprovidesevidenceforinteractiveeffectsbetweendelayedbedtimeandfamily concernfactorswhichsynergisticallyelevatedtheriskofchildhooddepression.This informationmayserveasapracticalguideforparentsandschoolteachersbyrecognizing that an adequate bedtime schedule could serve as a preventive measure against depressioninchildren.
ß2011ElsevierLtd.Allrightsreserved.
*Correspondingauthorat:SchoolofPublicHealth,NationalDefenceMedicalCenter,No.161,Min-ChunE.Rd.,Sec.6,Taipei,Taiwan.
Tel.:+886287923100x18470;fax:+886287923147.
E-mailaddresses:fugong@ndmctsgh.edu.tw,fugong@mail.ndmctsgh.edu.tw(F.-G.Lin).
ContentslistsavailableatSciVerseScienceDirect
Research in Developmental Disabilities
0891-4222/$–seefrontmatterß2011ElsevierLtd.Allrightsreserved.
doi:10.1016/j.ridd.2011.08.011
Depressioninadultsmayoriginatefromfactorsstemmingfromtheirchildhoodandfamilyenvironment,especiallyfor thosewhocomefromfamilieswithlowsocioeconomicstatus,familydisruption,parentaldivorceorahistoryofmental illness.Suchchildrenmaysufferfromahigherriskofdepressioninadultlife(Gilman,Kawachi,Fitzmaurice,&Buka,2002;
Melchior,Moffitt,Milne,Poulton,&Caspi,2007;Tracy,Zimmerman,Galea,Mccauley,&Stoep,2008).Currentstudiessuggest thatdepressivesyndromesinchildrenmayhaveanegativeimpactontheirdevelopment,schoolperformanceandfamily relationships and may be an important precursor to psychopathology in adulthood (Bhatia & Bhatia, 2007; Chang, Zauszniewski,Heinzer,Musil,&Tsai,2007).
Depressionisfoundtobeassociatedwithothermentaldisorderssuchasanxietydisorder,oppositionaldisorder,substance usedisorderandconductdisorder(Fergusson,Boden,&Horwood,2010).TheTaiwaneseNationalsurveydatain1999indicated that30.5%ofadolescentsbetweentheagesof12and18experienceddepressivesymptoms(Lueetal.,2010).Researchon measuringdepressivesymptomsconductedin2007revealedthatalmostaquarterofthechildren(24%)werefoundhaving depressivesymptoms,and4.2%ofthechildrenwerereportedwithmajordepressivedisorders(Chang,Chen,&Lien,2007).
Riskfactorsforadolescentsandchildhooddepressionweredocumentedinvariousstudiesrevolvingaroundmultipleand complexissuesofpersonalcharacteristics,familyandschoolenvironments(Wu,Kao,Yen,&Lee,2007).Referringtopersonal characters,age,sex,bodyweightandsleepbehaviorwereallaccountedfor(Erickson,Robinson,Haydel,&Killen,2000);
whilethefamilialcomponentsofsocialeconomicstatus,familyclimate(Gilmanetal.,2002),parentingstylesandthe depressionhistoryoftheparents(Freres,Gillham,Reivich,&Shatte,2002)werealsomentionedinchildhooddepression research.Thefactorsconcerningschoolenvironment,peerrelationships,supportandacademicperformancewererelatedto childmentalhealthandtheoccurrenceofdepression(Linetal.,2008).Childrenwhohadrelativelyfewfriendstendedto experiencedepressivesymptoms(Schwartz,Gorman,Duong,&Nakamoto,2008).
Current studies on the relationships between shorter sleep duration and depression were mostly carried out on adolescentswhoappearedtobemoresophisticatedinarticulatingviewpointsandexpressingtheiremotionswhenbeing measuredbythescalescomparedtopre-adolescents(Dewaldetal.,2010;Gangwischetal.,2010;Mooreetal.,2009).
Nevertheless, Hudson had reported that elementary school-aged children with clinically anxious symptoms had significantlyshortersleepdurationthanacontrolgroup.However,fewpapershavetheirfocuscenteredonpre-adolescent children.Inthiswork,thevariablesassociatedwiththefactorsofpersonalcharacteristics,familyandschoolweremeasured.
Thegoalwastoevaluatetheeffectsofhowinsufficientsleepinteractswiththesefactorswithregardstodepression.
2. Methods 2.1. Participants
In2009atotalof36elementaryschoolsinMiaoLiCountywhorangovernmental-assistedafterschoolprogramswere contacted.Atotalof29schoolsagreedtojointhestudycomprisingof1643studentsingrades3through6whoattended after-schoolclasses.The29schoolslocatedevenlyatthewestsideof10morepopuloustownsintheCounty,whereasthe residual8townswereprimarilylocatedonthelesspopulousmountainareasoftheeastsideoftheCounty.Thisstudywas approvedbytheInstituteReviewBoardoftheTri-ServiceGeneralHospital,NationalDefenseMedicalCenter(Approval number:98-05-239).Parentswereinformedofthecontentofthestudyandgiventheopportunitytoconsentorrefusethe participationoftheirchildrenwhilechildrenwerealsoallowedtorefusetheirparticipationatanytime.Surveysandphysical measurementofheightandweightweretakenduringafter-schoolclasshoursbyscaleandheightmachinewithtrained assistants.Finally676questionnaireswerecompletedandanalyzed.
2.2. Measurementofdepressivesymptoms
The‘‘DepressionScreeningScaleforChildrenandAdolescents’’developedbyChang,Chen,etal.(2007)forTaiwanese childrenandadolescents,wasusedtomeasurethedepressionstatus.Thescalewasacombinedandmodifiedversionof variousothermeasurementmethods,similartothosefoundintheDiagnosticandstatisticalmanualofmentaldisorder,3rd and 4th editionpublished by the AmericanPsychiatric Association, Reynolds ChildDepression Scale(RCDS), Children’s DepressionInventorybyKovacs(CDI)andBeckDepressionInventory(BDI).Therewere34itemsinChang’sscalecoveringfive factors, theyare: (1) depression-dysphoric mood; (2) hopelessness; (3) low-self esteem; (4) somatic complaintsand psychomotor and (5) social-conduct-problems. Each item was scored 0–4and subjectswere askedto indicate how frequentlytheyhadexperiencedadescribedsituationduringthelast2weeks.Substantialinternalconsistencyofthescale wasindicated bya Cronbach’salphaof 0.94.Scores summatedgreater than33 wereclassifiedas possibledepression disorders.InChang’sstudy,theverysamescalewasusedtotest hospitaloutpatientsamplesand95%oftheclinically diagnosedpatientswerecorrectlyclassifiedbythisscreeningscale.
2.3. Measurementsofbedtimeandgetuptime
Childrenwereaskedabouttheirusualbedtime and wake-uptime duringweekdays.Nighttime sleepdurationwas calculatedfromthetwotimepointsofbedtimeandwake-uptime.Sleepdurationreportedinthestudieswasgenerally interpretedastimespentinbed.Bedtimewasfurthercategorizedintofourtimegroupstocalculateandcomparetheriskof
differentbedtimeperiodsfordepression:before9PM,between9PMand10PM,between10PMand11PM,andafter11PM.
Getuptimewasalsogroupedasbefore5AM,between5AMto6AM,between6AMto7AM,andafter7AM.
2.4. Measurementsofindependentvariables
Therewere39questionsinthequestionnairedesignedtouncovervariablesinvolvingpersonalcharacteristics,family concernandschoolconcern.Personaldemographiccharacteristicsconsistedofage,genderandbodymassindex(BMI) whichwascalculated asweight/height2 in kg/m2 units.Children’s BMIwerecategorized asnormal, underweightand overweightbasedonthecriteriaoftheDepartmentofHealth,Taiwan.NormalweightwasdefinedashavingaBMIlying between5and84thpercentileforageand sex;overweightwashavinga BMI85thpercentile,andunderweightwas definedashavingaBMI<5thpercentile.
Familyconcernvariablesweremeasuredwithself-stateditemstodefinethehousehold:livingwithbothparentsor single-parent.Aspreviouslyreported(Beesdo,Pine,Lieb,&Wittchen,2010;Rhee,Lumeng,Appugliese,Kaciroti,&Bradley, 2006),themeasurementscalesforchildren’sperceptionofparentingstylesandfamilyclimatewereprimarilydesignedfor useinassessingadolescents,andnotdesignedassuchforusewithyoungchildren.Hencerespondentsnormallyrequireda longertimetocompletethequestionnaireandparentalassistancewasoftenrequiredduringtheprocess.Inourstudy,the subjectswereelementaryschoolchildrenandmayhavehaddifficultiescompletingatime-consumingquestionnaireorfail tofullycomprehendthequestionsasked.WethereforeadoptedamoresimplifiedscaledesignedbyChang,Chen,etal.
(2007)wherechildrenwereaskedabouttheirperceptionoftheparentingstyletheyexperiencedathome:(1)authoritative, annotatedasrespectfulofchild’sopinionswithclearboundaries,(2)authoritarian,annotatedasstrictdisciplinarian(3) permissive,annotatedasindulgentwithoutdisciplineor(4)neglectful,annotatedasemotionallyuninvolved;whilefamily climatewascategorizedas(1)harmoniousor(2)notharmonious.InChang’sownstudy,thesimplifiedscalesplayedavital roleinrecognizingdepressioninelementaryschoolchildren.Thedepressionscoresforchildrenperceivedtoexperiencean authoritativeparentingstyle was significantlylower than theother groups.Meanwhile, children reportingto have a harmoniousfamilyclimate at home scored lower in thescale. Measurements for thechild–parent relationshipwere simplifiedwiththetwooptionsofbeing‘‘close’’or ‘‘notclose’’.Schoolconcern variablesweremeasuredwithsubject academicperformancedividedintothreecategories,thefirst,middleandlowerthird,bytheteachers’assessments.Peer relationshipsweremeasuredbyself-reportasbeing‘‘close’’or‘‘notclose’’.
2.5. Statisticalanalysis
Dataanalysis wasconductedusing PASWStatistics 18.0software.TheChi-squaretest wasusedforevaluatingthe associationbetweencategoricalvariablesanddepressionstatusandt-testandANOVAtestwereemployedforcomparisonof continuousvariablessuchasageandsleepingdurationingroupeddata.
Intheanalysisfordepressivesymptoms,theresultwasexhibitedintwomaingroupsemployingthesummatedscoresof 33asacut-offpoint:‘atrisk’(>33)ofdepressionand‘notatrisk’(33).Dichotomizeddepressionscoresofsinglebinary(0,1) outcomewasusedtofitthelogisticregressionmodelpvalue<0.05wasconsideredstatisticallysignificantinallteststhat wereconducted.Multivariateanalysiswascarriedouttoevaluatetheoddsratioforassociationsbetweenthepresenceofa potentialriskoraprotectivefactorfordepressionwitha95%confidenceinterval.
Covariatesincludedintheadjustedmultivariatemodelwerepersonalcharacteristicsofage,sex,BMI,sleepdurationand thefamily’sreceiptofgovernmentalassistance;followedbyfamilyassociatedvariablesofparentmaritalstatus,family climate, child–parent relationship and familyparenting style; and finally with the school concern variables of peer relationshipandacademicperformanceassessmentbyteachers.
Oddsratioofmeasured variablesfordepressionwerecomputedbyChi-squareintheunadjustedunivariate model;
whereasinmulti-variateanalysis,oddsratioswerecomputedwithlogisticregressionmodel.
Inordertoexaminethecombinedeffectsontheriskofdepressionattributabletothepresenceofbothlaterbedtimesand eitherpoorchild–parentrelationships,familyclimate,orpeerrelationships,therelatedfamily/schoolsignificantvariables fordepressionwerestratifiedonadelayedbedtime(10PMasthecutpoint)basisbyusinglogisticregressiontoidentifyand comparetheindividualandcombinedoddsratioontheriskofdepression.Expectedcombinedeffectsforthetwoindicators fromanadditivemodelarecalculatedas(ORvariable1+ORvariable2 1),andfromthemultiplicativemodelcalculatedas (ORvariable1ORvariable2)aspreviousdescribed(Andersson,Alfredsson,Kallberg,Zdravkovic,&Ahlbom,2005).ExpectedOR andobservedOR(ORvariable1 andvariable 2)oftwoindicatorsondepressionriskwerethencomparedandexaminedthe existenceofinteractioneffect.Inallofthetestscarriedoutinthisstudy,ap-valueof<0.05wasconsideredtobestatistically significant.
3. Results
3.1. Descriptionofthesample
Thisstudyincludedasampleof676childrenwithasexratioof48%boysand52%girlswhoattendedafter-schoolclasses.
Themeanagewasat9.68yearsrangingfrom7to12years(seeTable1).Amongthem,oneinthreechildren(32.1%)received
governmentfinancialsupportwhiletwooutoffivechildren(41.3%)wereconsideredhavinganabnormalBMI,eitherunder or overweight. We foundthat theaverage mean depression score was 18.40 (SD=16.60) and 18% of children were categorizedas‘atriskofdepression’’.
3.2. Bedtimeandwake-uptimeperiods
Inourstudy,averagenocturnalsleepdurationduringweekdayswasjustbelow9h(8.6h).AsshowninTable1,about 28.6%ofstudentswenttobedbefore9PM,mostofthem(48.1%)weresetbedtimebetween9PMand10PM,andlessthan oneinfourchildren(23.3%)wenttobedafter10PM.Onthegetuptime,about30%casesgetupbefore6PM,andmorethan two-thirds(69%)ofchildrengotupbetween6AMand7AM.
Table1
Demographiccharacteristicsofchildreninthestudy.
Variable N=676 % MeanSDa(range)
Depressionscore 18.4016.60(0–102)
33 555 82.1
>33 121 17.9
Childrenvariables
Age(years) 676 9.681.16(7–12)
Nocturnalsleepduration(h) 676 8.601.01(5–13)
Bedtime
9PM 193 28.6
9–10PM 325 48.1
10–11PM 114 16.9
>11PM 44 6.5
Getuptime
5AM 14 2.1
5–6AM 190 28.1
6–7AM 467 69.1
>7AM 5 0.7
Gender
Female 353 52.2
Male 323 47.8
BMI
Normal 397 58.7
Under/overweight 279 41.3
Governmentassistance
No 459 67.9
Yes 217 32.1
Familyvariables Motherethnicity
Taiwanese 519 76.8
Foreign 157 23.2
Householdtype
Bothparents 572 84.6
Singleparent 104 15.4
Child–parentrelationship
Close 505 74.7
Notclose 171 25.3
Parentingstyle
Authoritative 616 91.1
Notauthoritativeb 60 8.9
Familyclimate
Harmonious 229 33.9
Notharmonious 447 66.1
Schoolvariables Peerrelationship
Close 351 51.9
Notclose 325 48.1
Academicperformance
Firsttwothirds 465 68.8
Lastthird 211 31.2
Nocturnalsleepdurationwascalculatedfrombedtimeandwake-uptime,BMI:bodymassindexgroupedaccordingtoDepartmentofHealth,Taiwan,R.O.C.
a SD:standarddeviation.
b Notauthoritativecategoryincludedthreecombineddataofauthoritarian,permissive,neglectfulgroups.
3.3. Variablesassociatedwithdepression
BasedontheanalysisofmeasuredvariablesexhibitedinTable2,thepersonalcharacteristicsofage,gender,BMIand governmentassistancedidnotappeartohavesignificantrelationswiththesubject’sdepressionstatus.Table2exhibitsthe oddsratiosormeasuredvariablesfordepression.Inthemodel1,resultsfromunadjustedunivariateanalysisshowedthat childrenwithbedtimesof10–11PM,after11PMweremorelikelytobedepressivethanthatwithbedtimebefore9PM,For familyvariables, children with ‘‘not close’’ child–parent relationship were found to have a 2.11-fold greater risk of depression,meanwhilechildrenwhoexperienceda‘‘harmonious’’familyclimatewerealmosttwiceaslikely(1.89fold)to beprotectedfromdepressivesymptomsasopposedtothosewhodidnothavea‘‘harmonious’’familyclimate.Concerning theschoolvariables,theriskfordepressionwasnearlydoubled(1.99-fold)forthereferencegroupcomparedwithchildren thatconsiderthemselvestohavea‘‘close’’ peer-relationship.Inthemodel2,consideringthecovariatesinthelogistic regressionanalysis,avitallinkwasestablishedbetweennocturnalsleepdurationanddepressionshowedthatthosewith
Table2
Relationofthechildcharacteristicsassociatedwithdepressioninunivariateandmultivariateanalysis(N=676).
Characteristics Model1a Model2b Model3c
UnivariateOR MultivariateOR MultivariateOR
(95%CI) (95%CI) (95%CI)
Demographicvariables
Age – 1.07(0.89–1.28) 1.06(0.88–1.27)
Nocturnalsleepduration – 0.74(0.61–0.89)** –
Bedtime
9PM 1 – 1
9–10PM 1.17(0.70–1.95) – 1.15(0.67–1.97)
10–11PM 1.90(1.04–3.46)* – 1.89(1.00–3.60)*
>11PM 4.48(2.36–10.08)** 4.66(2.14–10.15)**
Getuptime
6AM 1 – 1
>6AM 0.93(0.60–1.43) – 0.94(0.59–1.49)
Gender
Female 1 1 1
Male 1.05(0.71–1.56) 1.07(0.70–1.64) 1.09(0.71–1.67)
BMI
Normal 1 1 1
Under/overweight 0.88(0.59–1.32) 0.91(0.60–1.39) 0.94(0.61–1.43)
Governmentassistance
No 1 1 1
Yes 1.26(0.84–1.90) 1.08(0.68–1.72) 1.11(0.70–1.78)
Familyvariables Motherethnicity
Taiwanese 1 1 1
Foreign 1.24(0.79–1.94) 1.26(0.77–2.04) 1.31(0.81–2.13)
Householdtype
Bothparents 1 1 1
Singleparent 1.37(0.82–2.30) 1.31(0.75–2.30) 1.44(0.81–2.54)
Child–parentrelationship
Close 1 1 1
Notclose 2.11(1.39–3.20)** 1.78(1.13–2.79)* 1.83(1.16–2.89)**
Parentingstyle
Authoritative 1 1 1
Notauthoritative 1.03(0.52–2.05) 0.75(0.35–1.57) 0.73(0.34–1.55)
Familyclimate
Harmonious 1 1 1
Notharmonious 1.89(1.20–2.98)** 1.86(1.16–2.98)* 1.90(1.18–3.07)**
Schoolvariables Peerrelationship
Close 1 1 1
Notclose 1.99(1.33–2.98)** 1.97(1.29–3.03)** 1.95(1.27–3.01)**
Academicperformance
Firsttwothirds 1 1 1
Lastthird 1.33(0.88–2.00) 1.17(0.75–1.83) 1.16(0.59–1.49)
aModel1–unadjusted.
b Model2–includestimeinbedforsleephabitwithdemographic,familyandschoolrelatedvariablesadjustedinequation.
cModel3–includesbedtime,getuptimeperiodsforsleephabitswithdemographic,familyandschoolrelatedvariablesadjustedinequation;CI, confidenceinterval;OR,oddsratio.
*p<0.05.
**p<0.01.
longernighttimesleepduration(timespentinbed)were0.74-foldlesslikelytobeatriskofdepression.Andtheotherthree significant variables in model 1 of child–parent relationship, family climate and peer-relationship were also found significantinmodel2,withdemographic,familyandschoolrelatedvariablesadjustedinequation,whichtheresultwas compatiblewithmodel1.Inthemodel3,adjustedoddsratioofdelayedbedtimes10–11PM,after11PMfordepressionwere 1.89,4.66foldsrespectivelythanwhosebedtimebefore9PM.Converselyavitallinkwasestablishedbetweennocturnal sleepdurationanddepressionshowedthatthosewithlongersleepdurationwere0.74-foldless likelytobeatriskof depression.Forfamilyvariables,childrenwith‘‘notclose’’child–parentrelationshipwerefoundtohavea1.78-foldgreater riskofdepression,meanwhilechildrenwhoexperienceda‘‘harmonious’’familyclimatewerealmosttwiceaslikely(1.86- fold)tobeprotectedfromdepressivesymptomsasopposedtothosewhodidnothavea‘‘harmonious’’familyclimate.
Concerningtheschoolvariables,theriskfordepressionwasnearlydoubled(1.97-fold)forthereferencegroupcompared withchildrenthatconsiderthemselvestohavea‘‘close’’peer-relationship.Theanalysisontherelationwithbedtime-getup timeanddepressionstatusindicatedthatabedtimeofafter10PMorlaterwasadecidingfactorfordepressionastherisk levelwas2.28-foldhigherthanthosewhosebedtimeswerebefore10PM.Bycontrastgetuptimedidnotseemtoforma cruciallinkwithdepression.
3.4. Synergisticeffectsofvariablesondepression
Fourfactorswerefoundtobesignificantlyassociatedwithdepressivestatusandtheywerepresentedbyindependently adjustedeffectsusingamultivariablelogisticmodel(seeTable2).Tofurtherassessifchildren’sdelayedbedtimeinteracted withtheotherthreesignificantriskfactors,variablesweredichotomouslystratifiedtoexaminetheindividualeffectsand jointeffectsundertheconditionofbedtimeafter10PMonchilddepression.Individualeffectofeachvariablewithbedtime andcomparisonbetweenexpectedjointeffectsandobservedjointeffectsofstratifiedvariablesondepressionwerethen performedwithlogisticregressionmethods.AsshowninTable3,theindividualeffectofbedtimeandfamilyclimateshowed amoderateassociationof1.77and1.83riskoddscomparedtothereferencegroup.Theexpectedjointeffectsofthetwo variableswere2.60-fold(1.77+1.83 1)undertheadditivemodelandwere3.24-fold(1.771.83)undermultiplicative modelcalculatedaspreviouslystated(Anderssonetal.,2005;Botto&Khoury,2001).Whereastheobservedjointeffectsof thetwovariableswere4.35-foldshowingahigherstrengthofassociationwithdepressionthantheexpectedjointeffectsin bothadditiveandmultiplicativemodel.Thecombinedfactorsofbedtimeafter10PMandchildrenwhodidnotperceivea harmoniousfamilyclimatedemonstratedasynergisticeffect.
Equally,theinteractioneffectbetweenchildren’sbedtimeandchild–parentrelationshipondepressionwasanalyzed.
Theindividualriskoddsofthetwovariablesondepressionwere1.67and2.04respectivelywhichwerecomparatively greaterthanthereferencegroup.Afteradjustingtheothervariables,theoddsratiosofexpectedjointeffectsofthetwo variableswas2.71(1.67+2.04 1)undertheadditivemodelandwas3.41(1.672.04)inthemultiplicativemodel.The observedjointeffectsofthetwovariablesonthedepressionriskincreasedto4.73-foldimplyingthatasynergisticeffect wasalsopresentinthesituationofthechildrenwithdelayedbedtimeofafter10PManddidnothaveaclosechild–parent relationship.Additionally,toexaminetheinteractioneffectbetweenbedtime,andpeerrelationship,theindividualoddsof twovariablesondepressionriskwere2.36and3.36,respectively.Theriskoddsondepressionofexpectedjointeffectsof thetwovariableswere4.72(2.36+3.36 1)withtheobservedjointeffectsbeing4.01hencethesynergisticeffectwasnot obvious.
Table3
Interactiveeffectsofchildrenbedtimeandrelevantsignificantvariablesondepression.
Variable1 Variable2 CrudeOR(95%CI) AORa95%CI)
Harmoniousfamilyclimate Delayedbedtimeafter10PM
Yes No 1 1
Yes Yes 2.10(0.88–5.01) 1.83(0.74–4.50)
No No 1.76(1.01–3.07)* 1.77(1.00–3.11)*
No Yes 4.14(2.23–7.70)** 4.35(2.30–8.24)**
Closechildren–parentrelationship Delayedbedtimeafter10PM
Yes No 1 1
Yes Yes 2.06(1.21–3.52)** 2.04(1.18–3.53)*
No No 1.85(1.10–3.14)* 1.67(0.96–2.88)
No Yes 4.94(2.60–9.43)** 4.73(2.38–9.39)**
Closepeerrelationship Delayedbedtimeafter10PM
Yes No 1 1
Yes Yes 3.58(1.87–6.83)** 3.36(1.73–6.53)**
No No 2.57(1.87–6.83)** 2.36(1.37–4.00)**
No Yes 4.39(2.33–8.27)** 4.01(2.09–7.70)**
a Adjustedforchildren,familyandschoolassociatedvariablesasshowninTable1.
* p<0.05.
** p<0.01.
4. Discussions
Inthisstudy,withmultivariateandstratifiedanalysisstrategies,wewereabletoassesstheindependentandthejoint modificationeffectsofsleephabitswithfamilyandschoolfactorsondepressionstatus.Thecloserelationshipbetweenlater bedtimeanddepressioninchildrenshowedinourresultswascompatiblewithpreviousstudies(Gangwischetal.,2010), andfurthermore children withadelayedbedtime after10 PMwerenotonly foundtobeassociatedwithdepression independentlybutalsooneenhancedfactorondepressionriskinconsideringtogetherwiththefamilyconcernvariables.
Mostpreviousstudiespresentedtherisksofdepressionassociatedwithfamilyand schoolfactorsforadolescentsand childrenasindependentvariables,thusmayhaveunderestimatedthejointseffectofthesevariables.Withtheanalysis,the jointeffectsofthedelayedbedtimeandnon-harmoniousfamilyclimateshowedmultiplicativeinteractionontheriskratio ofdepressionupto4.35-foldgreaterthanthereferencegroup;whilethejointeffectsofthedelayedbedtimeandnotclose child–parentrelationshipalsoshowedmultiplicativeinteractionof4.73-foldgreater.Theinteractedresultsoffamilyfactors andbedtimeondepressionmightindicatethatsleephabitsofelementaryschoolchildrenwerestilldependentonparents setbedtime.Children’semotionmightbeinfluencedbychild–parentinterrelationshipandfamilyclimateandledtolate bedtime.Thiswouldbesupportedbythedatathatdepressedchildrenreportedtohavehigherproportionnon-harmonious familyclimateandnot-closechild–parentrelationshipthanthenon-depressed.Thisalsorevealsthatfamilyfactorsoffamily climateandchild–parentrelationshipplayedimportantrolesinchildmentalhealth,especiallyinthosewhowiththesleep habitoflatebedtime.
Inourcases,childrenat4–6grades,withtherequirementtoattendschoolatregularscheduledtimearound7:30AM andthenormalclassendedat4:00PM,withattendingtheafterclassextendingto6PM,theyspenddaytimealmostin theschool.The nightsleepis theonly time forthemto restandsurelyan importantperiod fortheir physicaland psychologydevelopment.However, the mean timespent in bedof children in theinvestigation was 8.6handwas obviouslyshorterthan9.8hforchildrenagedbetween8–12aspreviouslyreportedbyIglowsteinetal.(2003)andthe averagesleepdurationof10.2hfor10-year-oldchildrenreportedbySzymczak,Jasinska,Pawlak,andZwierzykowska (1993).Ontimetoschoolmightbeonestressorforchildren,especiallyforthechildrenwithlatebedtime.Iglowstein et al.’s (2003) study reported that sleep duration had decreased for children born between 1974 and 1993 with progressivelylaterbedtimebutunchangedgetuptimeacrossdecades.Ourdatashowedthatchildrenwhowenttobed after10PMhadsignificantlyhigherdepressionriskthanthosebefore10PMregardlessofthegetuptimeintheadjusted logisticregressionmodel.Therefore,itimpliedthatthedelayedbedtimemediatedthesleeplengthofschoolchildrenand subsequentlyinfluencedthedepressionstatus.Studiesshowedthatchildrenandadolescentssetbedtimewerehighly correlatedwithdepressionsyndrome(Gangwischetal.,2010).TheTaiwaneseparent-setbedtimemaybeconsideredtoo lateforchildrenanditiscertainlyworthtakingintoaccountthepossiblepsychologicalimpactwhenitcomestosetting anadequatebedtime.
Analysisofthedemographicfactorsrevealedthatchildren’sage,genderandfamilyfactorssuchasfamilyclimateand child–parentrelationshipweresignificantlyrelatedtotheirnocturnalsleeptime.Boysspentashortertimeinbedof8.48h comparedtogirlsof8.71h(p<0.01)andthedurationwereshorterwitholderchildren.Childrenaged12hadtheshortest nighttimeinbedof8.11hcomparedwithothers(p<0.01)asshowninFig.1.Bothageandgenderweresignificantly correlatedwithnighttimesleepdurationwhichinturnimpliedthatageandgendermightbetheintermediateriskfactorsfor depression.
Previousstudiesprovidedevidenceinassociatingsleepdisorderwithchildren’sschoolperformance(Gozal,1998)and longersleepdurationwasalsolinkedwithbetterschoolperformance(Dewaldetal.,2010).Researchesonsleepphysiology showedthattruncatedsleepbroughtonbydelayedbedtimewasassociatedwithadelayedendogenousmelatoninonsetand delayeddaytimephasecircadianrhythm(Burgess&Eastman,2004;Yangetal.,2001).Delayedbedtimeresultedinbotha decreaseinself-ratedalertnessduringthedaytimeandadecreaseinstudents’academicperformance(Chung&Cheung, 2008;Clodore´ etal.,1987).Regularcircadianrhythmsofsleeppatternandanincreaseintotalsleeptimewillresultin
Fig.1.Relationofagebysexwithsleephoursinchildren.
improveddaytimefunctioning(Chung&Cheung,2008;Monk,Petrie,Hayes,&Kupfer,1994).Thirty-sixpercentofchildren hadadelayedbedtimelaterthan10PM.Theyshowedpooreracademicperformancethanchildrenwhowenttobedbefore 10PMwhereonly29.6%underperformacademically.Delayedbedtimeforchildrenwasassociatedwithdepressionand mightalsoinfluencetheiracademicperformanceviadisturbanceofcircadianrhythm.
Thesamplein thisstudywastakenfromafter-school classstudentsinrural areasofMiaoLi County.Someother confounding variableswere therefore not evaluated for, including urbanization. Sincethe study is crosssectional, causalityoranydirectionofcausalitycannotbedetermined,buttherearemanyinterestingpossibilitiesthatcouldbe exploredsuchas:islaterbedtimeandresultantshortersleepdurationreflectiveofinsomnia,asymptomofdepression, andthereforereflectstheseverityofdepression?Anddoeslackofsleepleadtodepressedmoodandirritabilitywhich functionstonegativelyimpactthechildren’sinterpersonalrelationshipswithparentsorclassmates?Thereexistsome potentialbidirectionalrelationshipsamongdelayedbedtime,familyfactorsanddepressionstatusinchildrenwhichmay needfurtherfollow-up.Inordertoreducetheriskofdepressioninchildren,wesuggestthatmoreconcernshouldbe placedonthechildhavingregularandsufficientsleepdurationaswellasimprovementsinthefamilyenvironment, especiallytherelationshipbetweentheparentsandthechild.Thesetwofactorsshouldhavethemaximumimpactonthe riskreduction.
Acknowledgements
Wewouldliketoacknowledgethecontributionsoftheteachersintheafterschoolclasswhoparticipatedinthisstudy.
AndspecialthankstoDr.ChangKao-Pinforhiskindlyauthorizeduseofthe‘‘DepressionScreenScaleforChildrenand Adolescents’’.
References
Andersson,T.,Alfredsson,L.,Kallberg,H.,Zdravkovic,S.,&Ahlbom,A.(2005).Calculatingmeasuresofbiologicalinteraction.EuropeanJournalofEpidemiology,20, 575–579.
Beesdo,K.,Pine,D.S.,Lieb,R.,&Wittchen,H.U.(2010).Incidenceandriskpatternsofanxietyanddepressivedisordersandcategorizationofgeneralizedanxiety disorder.ArchivesofGeneralPsychiatry,67,47–57.
Bhatia,S.K.,&Bhatia,S.C.(2007).Childhoodandadolescentdepression.AmericanFamilyPhysician,75,73–80.
Botto,L.D.,&Khoury,M.J.(2001).Commentary:Facingthechallengeofgene–environmentinteraction:Thetwo-by-fourtableandbeyond.AmericanJournalof Epidemiology,153,1016–1020.
Burgess,H.J.,&Eastman,C.I.(2004).Earlyversuslatebedtimesphaseshiftthehumandimlightmelatoninrhythmdespiteafixedmorninglightsontime.
NeuroscienceLetters,356,11511–11518.
Chang,H.J.,Zauszniewski,J.A.,Heinzer,M.M.,Musil,C.M.,&Tsai,W.C.(2007).Adaptivefunctioninganddepressivesymptomsinschool-agedchildren.Journalof AdvancedNursing,60,502–5122007.
Chang,K.,Chen,M.,&Lien,T.(2007).AsurveystudyoftheprevalenceofchildrendepressioninTaiwan.NTTUEducationalResearchJournal,18,29–72.
Chung,K.F.,&Cheung,M.M.(2008).Sleep–wakepatternsandsleepdisturbanceamongHongKongChineseadolescents.Sleep,31,185–194.
Clodore,M.,Benoit,O.,Foret,J.,Touitou,Y.,Touron,N.,Bouard,G.,etal.(1987).Earlyrisingordelayedbedtime:Whichisbetterforashortnight’ssleep?European JournalofAppliedPhysiologyandOccupationalPhysiology,56,403–411.
Dewald,J.F.,Meijer,A.M.,Oort,F.J.,Kerkhof,G.A.,&Bogels,S.M.(2010).Theinfluenceofsleepquality,sleepdurationandsleepinessonschoolperformancein childrenandadolescents:Ameta-analyticreview.SleepMedicineReviews,14,179–189.
Erickson,S.J.,Robinson,T.N.,Haydel,K.F.,&Killen,J.D.(2000).Areoverweightchildrenunhappy?:Bodymassindex,depressivesymptoms,andoverweight concernsinelementaryschoolchildren.ArchivesofPediatrics&AdolescentMedicine,154,931–935.
Fergusson,D.M.,Boden,J.M.,&Horwood,L.J.(2010).Structuralmodelsofthecomorbidityofinternalizingdisordersandsubstanceusedisordersinalongitudinal birthcohort.SocialPsychiatryandPsychiatricEpidemiology.
Freres,D.R.,Gillham,J.E.,Reivich,K.,&Shatte,A.J.(2002).Preventingdepressivesymptomsinmiddleschoolstudents:ThePennResiliencyProgram.International JournalofEmergencyMentalHealth,4,31–40.
Gangwisch,J.E.,Babiss,L.A.,Malaspina,D.,Turner,J.B.,Zammit,G.K.,&Posner,K.(2010).Earlierparentalsetbedtimesasaprotectivefactoragainstdepression andsuicidalideation.Sleep,33,97–106.
Gilman,S.E.,Kawachi,I.,Fitzmaurice,G.M.,&Buka,S.L.(2002).Socioeconomicstatusinchildhoodandthelifetimeriskofmajordepression.InternationalJournal ofEmergencyMentalHealth,31,359–367.
Gozal,D.(1998).Sleep-disorderedbreathingandschoolperformanceinchildren.Pediatrics,102,616–620.
Hudson,J.L.,Gradisar,M.,Gamble,A.,Schniering,C.A.,&Rebelo,I.(2009).Thesleeppatternsandproblemsofclinicallyanxiouschildren.BehaviourResearchand Therapy,47,339–344.
Iglowstein,I.,Jenni,O.G.,Molinari,L.,&Largo,R.H.(2003).Sleepdurationfrominfancytoadolescence:Referencevaluesandgenerationaltrends.Pediatrics,111, 302–307.
Lin,H.C.,Tang,T.C.,Yen,J.Y.,Ko,C.H.,Huang,C.F.,Liu,S.C.,etal.(2008).Depressionanditsassociationwithself-esteem,family,peerandschoolfactorsina populationof9586adolescentsinsouthernTaiwan.PsychiatryandClinicalNeurosciences,62,412–420.
Lue,B.H.,Wu,W.C.,&Yen,L.L.(2010).ExpressedemotionanditsrelationshiptoadolescentdepressionandantisocialbehaviorinnorthernTaiwan.Journalofthe FormosanMedicalAssociation,109,128–137.
Melchior,M.,Moffitt,T.E.,Milne,B.J.,Poulton,R.,&Caspi,A.(2007).Whydochildrenfromsocioeconomicallydisadvantagedfamiliessufferfrompoorhealth whentheyreachadulthood?Alife-coursestudy.AmericanJournalofEpidemiology,166,966–974.
Monk,T.H.,Petrie,S.R.,Hayes,A.J.,&Kupfer,D.J.(1994).Regularityofdailylifeinrelationtopersonality,age,gender,sleepqualityandcircadianrhythms.Journal ofSleepResearch,3,196–205.
Moore,M.,Kirchner,H.L.,Drotar,D.,Johnson,N.,Rosen,C., Ancoli-Israel,S.,etal.(2009).Relationshipsamongsleepiness,sleeptime,andpsychological functioninginadolescents.JournalofPediatricPsychology,34,1175–1183.
Rhee,K.E.,Lumeng,J.C.,Appugliese,D.P.,Kaciroti,N.,&Bradley,R.H.(2006).Parentingstylesandoverweightstatusinfirstgrade.Pediatrics,117,2047–
2054.
Schwartz,D.,Gorman,A.H.,Duong,M.T.,&Nakamoto,J.(2008).Peerrelationshipsandacademicachievementasinteractingpredictorsofdepressivesymptoms duringmiddlechildhood.JournalofAbnormalPsychology,117,289–299.
Szymczak,J.T.,Jasinska,M.,Pawlak,E.,&Zwierzykowska,M.(1993).Annualandweeklychangesinthesleep-wakerhythmofschoolchildren.Sleep,16, 433–435.
Tracy,M.,Zimmerman,F.J.,Galea,S.,Mccauley,E.,&Stoep,A.V.(2008).Whatexplainstherelationbetweenfamilypovertyandchildhooddepressivesymptoms?
JournalofPsychiatricResearch,42,1163–1175.
Wu,W.C.,Kao,C.H.,Yen,L.L.,&Lee,T.S.(2007).Comparisonofchildren’sself-reportsofdepressivesymptomsamongdifferentfamilyinteractiontypesin northernTaiwan.BMCPublicHealth7116.
Yang,C.M.,Spielman,A.J.,D’ambrosio,P.,Serizawa,S.,Nunes,J.,&Birnbaum,J.(2001).Asingledoseofmelatoninpreventsthephasedelayassociatedwitha delayedweekendsleeppattern.Sleep,24,272–281.