a meta-analysis of interventions that target children’s

A Meta-analysis of Interventions That TargetChildren’s Screen Time for Reduction

abstractBACKGROUND: Screen time, especially television viewing, is associ-ated with risk of overweight and obesity in children. Although severalinterventions have been developed to reduce children’s screen time,no systematic review of these interventions exists to date.

OBJECTIVE: This is a systematic review and meta-analysis of interven-tions targeting a reduction in children’s screen time.

METHODS: Effect sizes and associated 95% confidence intervals (CIs)were calculated by using a random-effects model. Heterogeneity tests,moderator analyses, assessment of bias, and sensitivity analyses wereconducted. Reliability was assessed with Cohen’s �.

RESULTS: The systematic search identified 3002 documents; 33 wereeligible for inclusion, and 29 were included in analyses. Most reportedpreintervention and postintervention data and were published in peer-reviewed journals. Although heterogeneity was present, no modera-tors were identified. Overall Hedges g (�0.144 [95% CI: �0.217 to�0.072]) and standard mean difference (SMD) (�0.148 [95% CI:�0.224 to�0.071]) indicated that interventions were linked with smallbut statistically significant reductions in screen time in children. Theresults were robust; the failsafe N was large, and the funnel plot andtrim-and-fill methods identified few missing studies.

CONCLUSIONS: Results show that interventions to reduce children’sscreen timehave a small but statistically significant effect. As the evidencebase expands, and the number of screen-time interventions increases,future research can expand on these findings by examining the clinicalrelevance and sustainability of effects, conducting a more thorough anal-ysis of effect modifiers, and identifying critical components of effectiveinterventions. Pediatrics 2011;128:e193–e210

AUTHORS: Dayna M. Maniccia, DrPH, MS,a Kirsten K.Davison, PhD, MS,a Simon J. Marshall, PhD, MA,b

Jennifer A. Manganello, PhD, MPH,a and Barbara A.Dennison, MDc,d

Departments of aHealth Policy, Management, and Behavior andcEpidemiology, School of Public Health, University at Albany,State University of New York, Albany, New York; bSchool ofExercise and Nutritional Sciences, San Diego State University,San Diego, California; and dDivision of Chronic Disease andInjury Prevention, New York State Department of Health, Albany,New York

KEY WORDSobesity, television, mass media, intervention, meta-analysis,program

ABBREVIATIONSSMD—standard mean differenceCI—confidence interval

Dr Maniccia contributed to the study conception and design;data collection, analysis, and interpretation; and manuscriptwriting. Dr Davison contributed to the study design andmanuscript writing and revision. Dr Marshall contributed to thestudy design, statistical analysis design, and manuscriptrevision. Drs Manganello and Dennison contributed to the studydesign and manuscript revision.

www.pediatrics.org/cgi/doi/10.1542/peds.2010-2353

doi:10.1542/peds.2010-2353

Accepted for publication Mar 22, 2011

Address correspondence to Dayna M. Maniccia, DrPH, MS,Department of Health Policy, Management, and Behavior, Schoolof Public Health, University at Albany, 1 University Place,Rensselaer, NY 12144. E-mail: [email protected]

PEDIATRICS (ISSN Numbers: Print, 0031-4005; Online, 1098-4275).

Copyright © 2011 by the American Academy of Pediatrics

FINANCIAL DISCLOSURE: The authors have indicated that theyhave no personal financial relationships relevant to this articleto disclose.

REVIEW ARTICLES

PEDIATRICS Volume 128, Number 1, July 2011 e193 by guest on October 4, 2021www.aappublications.org/newsDownloaded from

pediatrics.aappublications.org/

Although the American Academy of Pe-diatrics recommends no more than 2hours/day of screen time (watchingtelevision or videos/DVDs, playingvideo or computer games, and using acomputer for purposes other thanschool work) for children aged 2 yearsold and older,1,2 47% of children aged 2to 15 years spend 2 or more hours/dayusing screen media.3 A recent study4

found that approximately two-thirds ofa national sample of 3-year-old chil-dren watched �3 hours/day of televi-sion and were exposed to more than 5additional hours of indirect (ie, in thebackground) television daily. Contraryto the American Academy of Pediatricsrecommendations, 33% of childrenaged 6 years and younger5 and 71%of children aged 8 to 18 years6

have televisions in their bedrooms, acharacteristic associated with in-creased viewing.5–7

Time spent watching television is asso-ciated with a number of negativehealth behaviors and outcomes amongchildren, including overweight,8–11 ir-regular sleep,12 insufficient consump-tion of fruits and vegetables,13 and dis-ordered eating.14,15 Excessive televisionviewing has detrimental effects onprosocial behaviors, such as spendingtime with parents and siblings, doinghomework, or engaging in creativeplay,16 and is linked with getting lowergrades, getting into trouble, and feel-ing sadness and boredom.6 In addition,children who watch more televisionmay be exposed to more content thatnegatively influences healthy behavior.In particular, television and video-game violence is associated with vio-lence and aggression4,17 in thought anddeed, and exposure to tragedies anddisasters on television negatively im-pacts children’s mental health.18

Before1985,whenDietz andGortmaker19

published their seminal article linkingtelevision viewing toobesity, themajorityof the research onmedia effects on chil-

dren focused on the impact of televisioncontent ormedia policy on children’s be-havior.20–27, 28, 29–41 Recently, researchhas included developing and testing in-terventions that attempt to reducechildren’s screen time.42–47 Althoughmeta-analysis has been used to studythe relationship between media useand weight and physical activity,11 todate no systematic review and meta-analysis of interventions to reducechildren’s screen time has been com-pleted. In this article we present theresults of a systematic review andmeta-analysis conducted to identifyand examine the effectiveness of inter-ventions to change children’s screentime. The broader behavior of screentime was selected because, althoughtelevision use has remained relativelystable for more than 50 years,48 totalscreen time has increased over thepast 10 years,6 and the American Acad-emy of Pediatrics recommends limit-ing children’s total screen time.2,49

METHODS

Study Identification

Studies were identified with a system-atic search of research databases, re-view of the table of contents of journalsnot included in searchable databases,review of the reference lists of rele-vant publications, and a search of theNational Institutes of Health’s Com-puter Retrieval of Information on Sci-entific Projects (CRISP) funding data-base (currently “NIH RePORT”).

Between November 24 and December6, 2008, 8 databases available throughthe State University of New York at Al-bany Library were searched for thewords “television,” “media use,” “rec-reational media,” and “screen time,”combined with “trial,” “program,” “in-tervention,” and “experiment” (see Ap-pendix 1 for a sample search strategy).Search results were limited to workspublished after 1985 (the year Dietzand Gortmaker19 described the associ-

ation between children’s televisionviewing and weight). The Web-basedCochrane Library50 and the Centre forReviews and Dissemination51 data-bases also were searched using theterms listed above. The table of con-tents of 2 journals that have only re-cently been indexed in Medline, Cyber-psychology and Behavior and PediatricExercise Science, also were searchedby accessing the journal’s Web site.

The reference lists of the CochraneCollaboration reviews of obesity inter-ventions52,53 and articles selected forinclusion in the meta-analysis were re-viewed for eligible references. In an at-tempt to minimize the impact of thefile-drawer problem,54 the Academy ofMedicine’s Gray Literature Report da-tabase; the National Institutes ofHealth’s funding database, CRISP; andthe Web site worldwidescience.orgwere searched. Researchers identifiedin the CRISP search were contactedand asked to provide informationabout their National Institutes ofHealth–funded project if they believedit met the meta-analysis inclusioncriteria.

Identification of Eligible Studies

Articles identified by the systematicsearchwere screened for eligibility us-ing a 2-step process. First, referenceswere identified as eligible for full re-view if the title or abstract-statedscreen time was measured and tar-geted for change. If no abstract wasavailable, the complete reference wasreviewed. Second, articles eligible forfull review were screened to deter-mine eligibility for inclusion in themeta-analysis using the PICO (popula-tions, interventions, comparisons, andoutcomes) framework.55 A study waseligible for inclusion if it met all of thefollowing criteria: (1) described an in-tervention or program to change be-havior in children between the ages of0 and 18 years; (2) outlined the results

e194 MANICCIA et al by guest on October 4, 2021www.aappublications.org/newsDownloaded from

of an intervention to reduce screentime; (3) compared a nontreatmentcontrol, comparison group, or prein-tervention period with an interventiongroup or period; (4) included screentime (watching television or videos/DVDs, playing video or computergames, and using a computer for pur-poses other than school work alone orin combination) as an outcome vari-able; and (5)measured television view-ing alone or with any combination ofvideo viewing, computer time, or video-game use. Only articles published inEnglish were eligible for inclusion inthe meta-analysis.

Data Extraction and Coding

Information extracted from each arti-cle included sample characteristics(age, race/ethnicity, gender, and studylocation), intervention setting (home,school, or other), theoretical frame-work, intervention components (eg,goal setting, behavioral monitoring,and policy change), and behavior tar-geting (see Table 2). Study designinformation extracted included sam-pling and group-assignment proce-dures, timing of assessment, type ofcomparison group, and whether a val-idated instrument had been used tomeasure the outcome (see Table 3). Aninstrument was considered valid if theauthor stated that the instrument wasvalidated or on the basis of a previ-ously validated instrument. Whetherthe study targeted high-risk individu-als also was recorded; studies target-ing high-risk individuals only includedparticipants with high BMI and/or ex-cessive screen-time behavior. Thesample size at group assignment andeach assessment point and the num-ber of subjects included in the analysisalso were recorded. Finally, informa-tion about study outcomes, includingmeans and associated SDs and meanchange frombaseline to posttest, wereextracted for use in calculating effectsizes. Data were extracted using a

standard data extraction instrumentdeveloped specifically for this study.The instrument was based on theCommunity Guide methods and instru-ment,56–60 other review instru-ments,55,61–64 and accepted methodolo-gies.65,66 A copy of the data collectiontool and associated code book areavailable on request.

Statistical Analysis

Whenever possible, preinterventionand postintervention means (SD) wereused to calculate the study effect size.When unavailable, postinterventionmeans (SD), mean change in eachgroup, or adjusted differences afterthe intervention were used to calculatethe study effect size. If an exact P valuewas not provided, a conservative ap-proach of using an estimate closest tothe significance level provided was used(eg, if P � .01 was provided, P � .009was used to calculate the effect size).63,67

Given that the intent of the meta-analysis was to determine the overalleffectiveness of the interventions, amean study effect size was calculatedwhen multiple effect-size statisticscould be calculated55,68 (for example,when data were presented from sev-eral subsamples within a study orwhen multiple screen-time measureswere presented). When key informa-tion was missing from an article, aswas the case in 19 of 31 eligible stud-ies, 3 attempts weremade to reach thecorresponding author via e-mail be-fore using the available data or elimi-nating the study from the analysis.

Because of the subjectivity and the dif-ficulty with assessing the overall qual-ity of each study (because of insuffi-cient information), a single index ofstudy quality was not included in theanalyses. Instead, study characteris-tics that are reflective of study quality(eg, number of groups, random as-signment to groups) were examinedas potential moderators, as outlined in

greater detail below. An assessment ofthe heterogeneity of effect sizes wasconducted to confirm the appropriate-ness of a random-effects model. Sensi-tivity and publication bias analysesalso were conducted to improve accu-racy and assess the robustness of theresults.

Reliability Analyses

A random 25% subsample of articles(n � 8) eligible for inclusion in themeta-analysis was reviewed by a sec-ond reviewer to assess coding reliabil-ity. � was calculated by using Stata9.2 (Stata Corp, College Station, TX) todetermine coder agreement beyondchance.69,70 The percentage of agree-ment between coders also was calcu-lated. All items included in the analy-ses had at least an 80% coderagreement or a � value of �0.70. A �value of �0.60 has been deemedgood69,71 and 0.80 has been deemed ac-ceptable in many situations.70 Any dis-agreements were settled via consen-sus. After data entry, all studies werereviewed a second time to verify dataextraction and entry.

Effect-Size Calculations

Data were entered into and analyzedwith Comprehensive Meta-analysis 2(Biostat, Englewood, NJ).72 Two mea-sures of effect, the SMD and Hedges g,and associated 95% confidence inter-vals (CIs) were calculated. The datawere coded such that a negative effectsize indicated a greater reduction inscreen time in the interventiongroup relative to the control or com-parison group or preintervention pe-riod. A generally accepted criteriafor effect-size magnitude was ad-opted (0.2 � small; 0.5 � medium;0.8 � large).55,73,74

A random-effects model was used.55,75

Separate models were calculated forstudies that measured the amount ofscreen time during the interventionperiod (n � 5) and studies that re-

REVIEW ARTICLES



ported postintervention data (n� 27).Three studies were included in bothdata sets. When datawere provided formultiple time points after the interven-tion (eg, immediately after and 6months after the intervention), datafrom the time pointmost proximal to theintervention period were included in theanalyses. At least 10 studies are re-quired formoderatoranalyses tobecon-ducted.55 Because too few studies re-ported data collected during theintervention period (n � 5) moderatoranalyseswere conducted only with stud-ies that reported postintervention data.

Heterogeneity and ModeratorAnalyses

To assess heterogeneity and identifypossible moderator variables, effect-size estimateswere calculated for sub-groups and associated CIs were exam-ined for overlap. The between-groupsQ test, with a significance level of P�

.05, was used to assess significant dif-ferences between subgroups. Poten-tial moderators were determined apriori68,76 and included age, gender, ra-cial distribution (up to 50% nonwhite,�50% nonwhite), country, type of data(adjusted, raw), number of studygroups, population risk, use of a televi-sion control device, outcome (screentime, television alone), interventionsetting (school, home, other), and the-oretical model. Because the Q statisticdoes not provide an assessment of themagnitude of heterogeneity,77 the I2

value68,78–80 was calculated to assessthe proportion of variance that re-flects real difference in effect sizes.68,79

Publication Bias

Publication bias was assessed using 3techniques, the funnel plot,81 failsafeN,82 and the Duval and Tweedie trim-and-fill method.83,84 In the absence ofpublication bias, the distribution of ef-fect sizes in the funnel plot are sym-metrical and take on an inverted fun-nel shape.63,85 The failsafe N or file-

drawer number estimates the numberof studies that would need to be in-cluded in the meta-analysis to changethe overall results.55,82 The Duval andTweedie trim-and-fill method assumesthat the most undesirable studies aremissing.86 An asymmetric appearanceof many missing studies suggest pub-lication or small-sample bias.55

Sensitivity Analyses

Several sensitivity analyses were con-ducted to assess the robustness of theresults: the SMD and Hedges g werecompared; the impact of including ad-justed data was assessed; and the im-pact of individual studies was deter-mined.55 To assess the impact ofincluding studies that provided an ad-

justed value for changes in televisionviewing behavior or postinterventionmeans, the effect size was calculatedwith and without studies that providedadjusted data (n � 4). The overall ef-fect size also was calculated repeat-edly, with 1 study excluded each timeto determine whether the overall ef-fect size was strongly influenced by 1particular study.

RESULTS

Systematic Literature Search

Fourteen databases were searched,resulting in 3002 potential studies.Thirty-three studies were eligible forinclusion in the meta-analysis (Fig 1).Four studies were excluded from the

Database searches and number of potentially eligible references retrieved from each

Centre for Reviews and Dissemination (n = 19) Cochrane library (n = 158) CRISP (n = 53) Dissertation and theses (n = 33) Medline - Ebsco (n = 141) Medline - PubMed (n = 420) National Academy of Medicine Grey Literature (n = 33) PapersFirst and Proceedings First (n = 73) PsychInfo (n = 723) Science Direct (n = 11) Scirus (n = 698) Social Sciences Abstracts (n = 545) Social Work Abstracts (n = 68)

Potentially relevant references (duplicates removed)

n = 2539

Other sources of potentially relevant references (No. of references)

Personal files (n = 1) Reference list (n = 4) WorldWideScience.org (n = 458)

References eligible for title and abstract review

n = 3002

Excluded on basis of title and abstract review

n = 2951 Included in full

reviewn = 51 Reasons for excluding articles from the analyses

Case study design (n = 2) Did not describe an intervention (n = 7) Targeted adults (n = 1) Did not measure screen time (n = 2) Data presented elsewhere (n = 4) Author stated did not meet criteria (n = 2)

Eligible for inclusion in analyses

n = 33 Excluded from analyses

Data not available (n = 2) Two cross-sectional samples (n = 1) Outcome met criteria of <1 h viewing (n = 1)

Included in analyses n = 29

FIGURE 1The study-identification process.


TABLE1DescriptionofInterventionsEligibleforInclusionintheMeta-analysis

Author(Year)

InterventionName

InterventionDescription

SummaryofInterventionEffectivenessBasedonthe

Authors’Results

Interventionsincludedin

themeta-analysis

Angelbuer91(1998)

TheSwitchApparatus

Theinterventionconsistedoffamilycounseling,includinganeducationalcomponent,television

planning,anactivitiesmenuandacontractwithrewardsforbehavior.Participantswere

providedatelevision-controldevice.

Televisionviewingwasreducedfrom

�30h/wkduring

baselinetoslightlymorethan10h/wkafterthe

intervention.

Chinetal44(2008)

DOiT(DutchObesity

Interventionin

Teenagers)

Theinterventionconsistedofabehavioralcomponent(11lessonsforbiologyandphysical

educationclassestoraiseawarenessandfacilitatepositivehealthbehaviors)andan

environmentalcomponent(encouragingadditionalphysicaleducationclassesandchanges

intheschoolcafeteria).

Nosignificantchangeinscreen-viewingbehavior

occurred.

Dennisonetal42(2004)

BrocodiletheCrocodile

Theinterventionincreasedawarenessbyusingmediadiariesandincludedactivitiesinchild

carecenters.Childrenwereencouragedtoturnoffthetelevisionandwereactivelyinvolved

inidentifyingalternativestowatchingtelevision.

Significantdifferencesbetweengroups;theintervention

groupdecreasedtelevisionandvideoviewingwhilethe

controlgroupincreasedtimespentviewingtelevision

andvideos.

Eisenmannetal45(2008)SwitchWhatYouDo,Chew,

andView

Amultilevel(community,school,andfamily)interventiontomodifyphysicalactivity,screen

time,andnutritionbehaviors.Thecommunity-basedcomponentconsistedofacampaignto

increaseawarenessandknowledgeofpreventingchildhoodobesity.Theschool-based

componentincludedaschool-widekickoffeventandateacherpacketwithmaterialsfor

classroomdisplayanddistributiontothechildrenandideasforintegratingmaterialsinto

thecurricula.Thefamilycomponentincludedinformationonstrategiesforcreatingafamily

environmentsupportiveofhealthybehaviors.

Significantdifferencesinparent-reportedscreentime

immediatelyafterand6moaftertheinterventionwere

seenintheinterventiongroup.Changeinchild-

reportedscreentimewasnotsignificantbutwas

loweraftertheintervention.

Epsteinetal43(2008)

Televisionallowance

Weeklybudgetsforscreentimeweresetandbudgetsreducedby10%permonthuntilthe

budgetwas50%ofbaseline.Whenthebudgetwasreached,thetelevisionsetcouldnotbe

turnedon.Rewardswereearnedforstayingunderbudget.Parentsalsowereprovidedwith

tipstoreducesedentarybehavior.

Theinterventionresultedinastatisticallysignificantand

sustainedreductionintelevisionviewingand

computeruse.

Faithetal112(2001)

Televisioncycle

Duringtheintervention,televisionviewingwascontingentonpedalinganexercisecycleator

aboveaprescribedintensitylevel.

Amongthetreatmentgroup,television-viewingtime

significantlydecreasedfrombaselinetoweeks3to5,

andamongthecontrolgrouptelevisionviewing

increased.Duringthe12-wkperiod,televisionviewing

declinedinbothgroups,withgreaterdeclinesinthe

interventiongroup.

Fordetal92(2002)

Televisionallowance

Aclinicalinterventionthatprovidedfamilieswithbriefcounselingabouttheproblems

associatedwithexcessivemediaand15-to20-minutediscussionsaboutsettingtelevision

budgets.Parentsreceivedatelevision-controldevice(televisionallowance)tohelpbudget

televisiontime.

Bothgroupsreporteddecreasesinscreentime.

Differencesbetweengroupswerenotsignificant.The

familieswhousedthetelevision-controldevice

generallyreportedgreaterchanges.

Fosteretal47(2008)

Theinterventionincludedaschoolself-assessment,nutritioneducationandpolicy,social

marketing,andparentoutreach.SchoolstaffwereprovidedwithPlanetHealthandKnow

YourBodycurriculaandsupportingmaterials;theyalsoreceivednutritionandphysical

activitythemepacketstobeintegratedintoclassroomlessons,cafeteriapromotions,and

parentoutreach.Thefamilyoutreachcomponentconsistedofthe2-1-5challenge,during

whichchildrenwerechallengedtobelesssedentaryandmorephysicallyactive.

Timespentwatchingtelevisiondecreasedinthe

interventiongroupandincreasedinthecontrolgroup.

Golanetal93(1998)

Theinterventionusedparentsastheagentofchange.Parentsparticipatedingroupsession

originallyheldweekly,thenbiweekly,thanonceevery6wk(duringthistimefamiliesalso

attendfive15-minindividualsessions).Atallsessionsparentsweretaughttoalterthe

family’ssedentarylifestyle.

Noreductionintelevisionviewingtimeoccurred.

REVIEW ARTICLES



TABLE1Continued

Author(Year)

InterventionName



Authors’Results

Goldfieldetal46(2006)

Tokentelevision

Screentime(television/VCR/DVD)wascontrolledbyanelectronicdevice(tokentelevision).

Whentokenswereinsertedintothedevice,thetelevisionwasactivated.Childrenworea

pedometerthatprovidedthemwithfeedbackabouttheirphysicalactivity.Tokenswere

earnedbybeingphysicallyactive.

Screentimedecreasedintheinterventiongroupand

increasedinthecontrolgroup.Thedifferences

betweengroupsweresignificant.

Gortmakeretal94(1999)PlanetHealth(PowerDown)Aschool-basedinterventionthatincorporatedinterventionmaterialintomajorsubjectareas

andphysical-educationclasses.Materialsincludedcurriculumstandardssoskillsand

competenciesareusedtoconveythePlanetHealthmessages.Atotalof16corelessons

weredeliveredeachofthe2y.Anadditionallessondevelopeda2-wkcampaigntoreduce

televisionviewinginhouseholds(PowerDown).

Afteradjustingforbaselinecovariates,theNo.ofhoursof

televisionperdaywasreducedintheintervention

schoolscomparedwithstudentsincontrolschools.

Gortmakeretal95(1999)EatWellandKeepMoving

Program

Amulticomponentschool-basedinterventionthatconsistedofclassroommaterialsthatwere

integratedintomath,science,languagearts,andsocialstudiesclasses.Theintervention

alsoincludedactivitiesathometoinvolvefamilymembers.

Therewasanonsignificantreductionintelevisionand

videoviewing.

Harrisonetal96(2006)

SwitchOffGetActive

A10-lesson16-weekhealth-educationinterventionthatemphasized2keymessages:theneed

tominimizescreentimeandtheneedtoincreasephysicalactivity.

Therewasadecreaseinscreentimeintheintervention

groupbutpostinterventionviewingwasnot

significantlydifferentfromthecontrolgroup.

Jason97(1987)

Atoken-activateddevicewasattachedtothetelevisionset.Childrenhadtoearntokensby

participatinginprosocialactivities.Unusedtokenscouldbeexchangedfornontelevision

rewards(eg,money).

TheNo.ofhoursoftelevisionviewingdecreasedduring

andaftertheintervention.

Jasonetal98(1993)

TheSwitchApparatus

Adevicewasattachedtothetelevisiontocontrolelectricitytothesetandassociateddevices;

parentscontrolledturningtheswitchonandoff,therebycontrollingtelevisionviewing.

Thelockwaseffectiveinreducingtimespentviewing

television.

Jonesetal113(2008)

IMPACT(IncorporatingMore

PhysicalActivityand

CalciuminTeens)

Theinterventionusedlearningandenvironmentalreinforcementtoaffectbehaviorchange.It

includedahealthcurriculumthatincludedclassroomlessonsandbehavioraljournalism,

reportingrolemodelstories,aphysicaleducationprogram,andaschoolfoodservice

component.Theinterventionmaterialswereincorporatedintothehealthandphysical

educationcurricula.

Totalscreentimewassignificantlylowerinthe

interventiongroupcomparedtothecontrolgroup.

Kippingetal109(2008)

ActiveforLife(Year5)

ThisinterventionwasanadaptationandabbreviatedformoftheEatWellKeepMoving

intervention,whichconsistedof16lessonsonhealthyeating,increasingphysicalactivity,

andreducingtelevisionviewing.

Theinterventiongroupspentlesstimeonscreenviewing

activitiesthanthecontrolgroup,butthedifference

wasnotsignificant.

Maurielloetal114(2006)

HealthinMotion

Theinterventionwastailoredtoparticipant’sstageofchange.Participants’stageofchange

wasassessedandtheyreceivedfeedbackfromacomputerizedprogramtailoredtotheir

stage.ThecomputerprogrambeginswithanintroductiontotheHealthinMotionprogram

andthenproceedswithalternatingassessmentsandfeedbackabouttargetbehaviors.

Thetreatmentgroupreportedsignificantlylesstelevision

viewing.

McCanna110(1989)

Theinterventionconsistedofattachingatelevision-controldevice(token-actuatedtimertothe

television);inordertousethetelevision,tokenswereinsertedintothetimer.Participants

receivedtelevisiontokensforparticipatinginphysicalactivity.

Theinterventionwasmosteffectivefor1excessive

televisionviewer;forthebelow-averageviewers’

television-viewingrateswereminimallyaffected.

Nemetetal99(2005)

Childrenandparentswereprovidedcounselingandinvitedtolectures.Theintervention

includeddietarycounselingandatwice-weeklyexerciseprogram.Inaddition,participants

wereencouragedtoreducesedentaryactivityandreportwalkingorotherphysicalactivity

totheircoachesweekly.

Screentimedidnotchangesignificantlyforthe

interventionorcontrolsubjects.

Niemeyer100(1988)

BooksandBeyond

Theprogramwasdesignedtopromoterecreationalreadinganddiscriminatetelevision

viewing.StudentswereawardedprizesforreadingaspecifiedNo.ofpagesofbooks.The

interventionincorporatedhomeactivities,includingchartingtimespentreadingand

viewingtelevisionandlearningtoself-monitortelevisionviewingthroughacritical

television-viewingskillscurriculum.Parentswereprovidedmaterialsaboutdiscriminate

television-viewinghabitstoreadanddiscusswithchildren.

Therewasnosignificantdifferencebetweenthe

interventionandcomparisongroups.


TABLE1Continued

Author(Year)

InterventionName



Authors’Results

Novaetal101(2001)

Theinterventionoccurredinafamilypracticeoffice.Childrenandparentswereinterviewed

aboutactivityandeatinghabitsandgivenaspecificdietanddetailedguidelinesregarding

physicalactivity.Activeparentalcommitmentwasrequested.Follow-upclinicalvisitswere

performedregularlyduringtheintervention(at1,2.5,4,6,9,12,15,18,and24mo).During

follow-upvisitsparentsandchildrenreportedthechild’sbehaviorssincethelastvisit.

Nosignificantvariationsincomputerortelevisionuse

werenotedwithineachgroupfrom0to12mo.

Robinson102(1999)

Theinterventionincludedeighteen30-to50-minlessonsthatwereincorporatedintothe

standardcurriculumfollowedbyatelevisionturnoff.Parentswereprovidednewsletters

designedtomotivatethemtohelpchildrenstaywithintimebudgets.Thenewslettersalso

suggestedstrategiesforlimitingscreentime.Familiesweregivenatelevisionallowanceto

helpwithbudgeting.

Theinterventionsignificantlydecreasedchildren’s

televisionviewingcomparedwithcontrolchildren.The

interventiongroupchildrenalsoreportedsignificantly

greaterreductionsinvideo-gameusethancontrol

children.

Robinsonetal103(2003)

StanfordGEMS(GEMS

JewelsandSTART–

SistersTakingActionto

ReduceTelevision)

GEMSJewelsdanceclasseswereoffered5d/wk(classesstartedwithasnackfollowedby

homeworkperiodandthenthedancesessionsfollowedbydiscussionsaboutthe

importanceofdanceinthecommunityandculture).STARTconsistedof5lessonsdelivered

duringhomevisits.Television-viewinggoalsweresetandstrategiesforreachingthegoal

discussed.Familiesweregivenatelevisionallowancetohelpwithbudgetingand

newsletterstoreinforcethelessonsandprovideupdatesondanceclasses.

Comparedwithcontrolchildren,thetreatmentgroup

reportedlessmediauseandasignificantdecreasein

totalhouseholdtelevisionuseatfollow-up.


SMART(StudentMedia

AwarenesstoReduce

Television)Classroom

Curriculum

Theinterventionincludedbudgetingscreentimeandlimitingphysicalaccesstoscreenmedia.

Atotalof1830-to50-minclassroomlessonsplusweekly5-to10-minboosterswere

delivered.Thecurriculumconsistedof4sectionsfocusingontelevisionawareness,

televisionturnoff,developingskillstoresisttelevisionviewing,andhelpingothersdecrease

televisionviewing.Atelevision-controldevice(televisionallowance)wasprovidedtohelp

budgettelevision.Childrenwereprovidedincentivesformaintainingtheirweeklymediause

budget.

Comparedwithcontrolchildren,theinterventiongroup

significantlydecreasedweekdaytelevisionviewingand

weekdayandSaturdayvideo-gameplaying.Therewere

nosignificantdifferencesbetweengroupsinchangein

self-reportedweekdayorSaturdaytimespentplaying

onacomputer.

Salmonetal105(2008)

Switch-Play

Thebehavioralmodificationcomponentoftheinterventionincludednineteen40-to50-min

lessonsdeliveredintheclassroom.Thelessonsaimedtoincreaseawarenessofcurrent

behaviorandalternativesandthebenefitsofphysicalactivity.Childrencompletedaweekly

contracttoundertakeswitchingoff1televisionprogramperweekovera4-wkperiod.

Childrenintheinterventiongroupwatchedsignificantly

moretelevisiononaveragecomparedwiththoseinthe

controlgroupaftertheinterventionandat6and12

mofollow-up.Nosignificanteffectswereseenfor

electronicgameplayingorcomputeruse.

Segeetal106(1997)

Healthcareprovidersweregivena1-htrainingsessionthatincludedareviewoftheparent

materialstobehandedout,discussionsofthetheoreticalbackgroundofeachintervention,

andtheopportunitytodiscussimplementationdetails.Providersweretodistribute

informationcardstoparentsanddiscussthematerialonthecardwiththematthetimeof

achild’shealthvisit.

Therewasnosignificantchangeinweekdaytelevision-

viewinghabitsandatrendtowardreductionsin

weekendtelevisionviewing(theinterventiongroup

wasslightlymorelikelytoreportreductionsin

weekendtelevisionviewingcomparedwiththecontrol

group).

Simonetal107(2006)

ICAPS(Intervention

CenteredonAdolescents’

PhysicalActivityand

SedentaryBehavior)

Theinterventionwasamultilevelprogramthatincludedaneducationalcomponentfocusing

onphysicalactivityandsedentarybehaviors.Theinterventionalsoincludednew

opportunitiesforphysicalactivityduringandaftertheschoolday.

Interventionstudentshadagreaterreductionovertime

oftelevisionandvideoviewingthancontrolsubjects.

Weintraubetal111(2008)SPORT(TheStanfordSports

toPreventObesity

RandomizedTrial)

Theinterventionincluded4(originally3)d/wkofcoedsoccer.Oneday/weekwasagameand

theotherdayspractice;sessionsincludeda2.25-h-longhomeworkperiodfollowedbyan

activityperiod.

Inconsistentresultswereseenforscreentime.

REVIEW ARTICLES



analyses; 1 study was excluded be-cause it presented data on 2 cross-sectional samples,87 1 study was ex-cluded because the outcome wasbinary (�1 hour/day of television)88

compared with all the other studiesthat presented continuous data (eg,hours per week of screen time), and 2studies were excluded because datawere unavailable.89,90 Twenty-ninestudies were included in the analyses.Although 1 of the predefined exclusioncriteria was non-English language, noarticles written in a language otherthan English were identified. Fifteenauthors were sent e-mails requestingadditional information, and all but 1 re-sponded. Eight authors provided data,4 authors were unable to provide thedata requested, and 2 authors statedthat their work did not meet the studyinclusion criteria. Search of the CRISPdatabase identified 53 records that in-cluded the search terms. Only 4 ofthese studies were eligible for inclu-sion; 1 author responded to the re-quest for additional information stat-ing that data were not available at thetime. The vast majority (n � 26) ofstudies were published in peer-reviewed journals, and 3 studies weredoctoral dissertations.

Sample and InterventionCharacteristics

Table 1 briefly describes the interven-tions eligible for inclusion in the meta-analysis and summarizes the studyoutcomes specifically related toscreen time (n� 33). Table 2 providesadditional information about the sam-ple and intervention characteristicsfor interventions included in the anal-yses (n � 29). More than one-half ofthe interventions included in the anal-ysis (18 of 29) were theory based (so-cial cognitive theory was used mostfrequently). The school was the mostcommon intervention setting (13 of29), followed by the home (8 of 29).Many programs included both parentsTA

BLE1Continued

Author(Year)

InterventionName



Authors’Results

Interventionsexcludedfrom

theanalysesbecause

oftheunavailabilityof

data

Epsteinetal89(1995)

Theinterventionconsistedofreinforcingchildrenforcertainbehaviors.Allgroupsreceived

writteninformationonthepositiveeffectsofincreasedphysicalactivityandthenegative

effectsofsedentarybehaviors.Participantsinthesedentarybehaviorgroupwere

reinforcedfordecreasingtheamountoftimespentinsedentaryactivitiesandparticipants

intheexercisegroupwerereinforcedforincreasingphysicalactivity.

Informationaboutchangesinsedentarybehaviorswas

notavailable.

Epsteinetal90(2000)

Participantsinthedecreasedsedentaryactivitygroupwerereinforcedforreducingsedentary

behaviors,includingwatchingtelevisionandvideotapesandplayingcomputergames.

Targetedsedentarybehaviorsshowedasignificant

decreasefrombaselineat6and24mo.

Johnsonetal87(2005)

HealthyHabits

Astatewideinterventioninwomen,infants,andchildrenclinics.Theinterventionconsistedof2

moduleseachconsistingofbackgroundmaterials,stafftrainingmaterials,posters,

interactivehandoutsforclients,bookmarks,vouchers,coloringmaterials,anddetailed

plansforgroupsessionsandothersupportivematerials.

Betweenthebaselineandthe6-mosurveys,the

proportionoffamilieswhometrecommendationsfor

televisionviewingincreased.

Johnstonetal88(2006)

HealthySteps

Parentsreceivedparentingclasses,developmentalandbehavioraladvice,riskfactor

screening,postnatalhomevisits,telephonesupport,developmentalassessments,andthe

ReachOutandReadliteracyprogram.

Comparedwiththecontrolgroup,parentsinthe

interventiongroupwerelesslikelytoallowtheir

childrenmorethan1hofdailytelevisionviewing.


TABLE2SampleCharacteristicsandInterventionComponentsforStudiesIncludedintheMeta-analysis

Author(Year)

SampleCharacteristics

InterventionCharacteristics

n;AgeRange,y;Nonwhite,

%;Male,%

Location

TargetsaHigh-Risk

Group

SettingTelevision-Control

DeviceUsed

TheoryUsedin

Program

Intervention

Components

Behavior(s)TargetedforChange

Angelbuer91(1998)

6;5–11;unclear;�50%

UnitedStates

Yes,screen

Home

Yes

Other

B,C,INFO,M,R

Television

Chinetal44(2008)

821;12–18;unclear;

�50%

Netherlands

NoSchool

NoTRA/TPB,otherCU,G,INFO,M,PA

ST,consumptionofSSB,dietary

behavior,activecommutingto

school

Dennisonetal42(2004)

77;�5;

�50%;�50%

UnitedStates

NoSchool

NoINFO,R

ST,snackingwhilewatching

television

Eisenmannetal45(2008)

20;5–11;unclear;

�50%

UnitedStates

NoHome

NoEST

INFO,M,R

ST,PA,dietarybehavior

Epsteinetal43(2008)

67;5–11;

�50%;�50%

UnitedStates

Yes,weight;yes,screen

Home

Yes

B,INFO,R

STFaithetal112(2001)

10;5–11;unclear;

�50%

UnitedStates


Home

NoOther

Contingent

Television(contingentonPA)

Fordetal92(2002)

28;5–11;

�50%;�50%

UnitedStates

NoOther

Yes

SCT

B,INFO

ST,eatingbreakfastanddinnerwith

televisionon,PA

Fosteretal47(2008)

689;5–11;�50%;�50%

UnitedStates

NoSchool

NoOther

INFO

Television,dietarybehaviors,PA

Golanetal93(1998)

50;5–11;unclear;

�50%

Israel

Yes,weight

Unclear

NoINFO

Television,nutritionalbehaviors,PA,

eatingwhilestanding,infrontof

television,whenstressed,while

performingotheractivities

Goldfieldetal46(2006)

30;5–11;unclear;

�50%

Canada


Home

Yes

Other

R,M

ST,snackintakewhilewatching

television,PA

Gortmakeretal94(1999)

1225;12–18;�50%;�50%

UnitedStates

NoSchool

NoSCT,BCT

CU,G,INFO,M

Television,nutritionalbehaviors,PA

Gortmakeretal95(1999)

336;5–11;�50%;�50%

UnitedStates

NoSchool

NoCU,INFO

Television,nutritionalbehaviors,PA

Harrisonetal96(2006)

312;5–11;unclear;�50%

Ireland

NoSchool

NoSCT

B,CU,G,INFO,M

ST,PA

Jason97(1987)

2;12–18;

�50%;�50%

UnitedStates

Yes,screen

Home

Yes

RTelevision

Jasonetal98(1993)

2;5–11;�50%;�50%

UnitedStates

NoHome

Yes

BST

Jonesetal113(2008)

606;12–18;

�50%;�50%

UnitedStates

NoSchool

NoSCT,TTM

CU,INFO

ST,PA

Kippingetal109(2008)

323;5–11;unclear;�50%

England

NoSchool

NoSCT,TTM

CU,INFO,M

ST,nutritionalbehaviors,PA

Maurielloetal114(2006)

655;12–18;

�50%;�50%

UnitedStates

NoUnclear

NoSCT,BCT

Stage

Television,nutritionalbehavior,PA

McCanna110(1989)

5;5–11;�50%;�50%

UnitedStates

NoHome

Yes

TTM

Television,alternativeactivities

(homework,reading,playing,

chores,sports)

Nemetetal99(2005)

46;5–11;unclear;

�50%

Israel

Yes,weight

Unclear

NoINFO,PA

ST,nutritionalbehavior,PA

Niemeyer100(1988)

235;notprovided;

unclear;unclear

UnitedStates

NoSchool

NoB,M

Television,reading

Novaetal101(2001)

140;5–11;unclear;�50%

Italy

Yes,weight

Other

NoINFO

ST,PA,nutritionalbehavior

Robinson102(1999)

192;5–11;unclear;�50%

UnitedStates

NoSchool

Yes

B,CU,INFO,M

ST,eatingmealsandsnackingin

frontofthetelevision


60;5–11;

�50%;�50%

UnitedStates

Yes,weight

Other

NoSCT

B,G,INFO,PA

ST,eatingbreakfastanddinnerwith

thetelevisionon,nutritional

behaviors,PA


181;5–11;unclear;�50%

UnitedStates

NoSchool

Yes

SCT

B,CU,INFO,M,R

STSalmonetal105(2008)

119;5–11;unclear;�50%

Australia

NoSchool

NoSCT

B,C,CU,INFO,M,R

ST,PA

Segeetal106(1997)

243;

�5;unclear;

�50%

UnitedStates

NoOther

NoSCT,BCT

INFO

Television

Simonetal107(2006)

954;12–18;unclear;

�50%

France

NoOther

NoINFO,PA,Policy

ST,PA

Weintraubetal111(2008)

21;5–11;

�50%;unclear

UnitedStates

Yes,weight

School

NoEST

PAST,PA

Theoriesusedinprogramdevelopment:BCT,behavioralchoicetheory;EST,ecologicalsystemstheory;SCT,socialcognitivetheory;TRA/TPB,theoryofreasonedaction/theoryofplannedbehavior;TTM,transtheoreticalmodel.Interventioncomponents:

B,budget;C,contract;Contingent,televisionusecontingentonanotherbehavior;CU,incorporatedintocurriculum;G,goalsetting;INFO,provisionofinformation;M,monitoringofbehavior;PA,physicalactivityopportunities;Policy,studyincludeda

policychangecomponent;R,rewards/reinforcements;Stage,theinterventionisbasedontheindividual’sstageofreadiness.Targetsahigh-risksample:Yes,weight,theinterventiontargetsoverweightchildren;Yes,screen,theinterventiontargets

childrenthatexceedascreentimerecommendations.Behavior(s)targetedforchange:ST,screentime;television,televisiononly;SSB,sugar-sweetenedbeverages;PA,physicalactivity;nutritionalbehaviors,dietaryhabits(eg,fatandcalorieintake,

fruitandvegetableconsumption).

REVIEW ARTICLES



and children in the intervention. With re-gard to intervention components, ap-proximately one-third (9 of 29) of the in-terventions facilitated behavior changeby controlling the environment with atelevision-control device. Many interven-tions facilitated behavior change by set-ting goals and planningmedia use; oftenchildren participated in this process. An-other common feature of the interven-tions was a behavioral contract in whichchildren agreed to a specified amount oftime in front of a screen. Often, a rewardwas provided if screen-time targetswere met. Several interventions in-cluded increasing awareness by havingchildren monitor and record their ownscreen time. Finally, only 1 interventionmade television viewing contingent onphysical activity.

The primary outcome in themajority ofinterventions (19 of 29) was screentime (television, video/DVD, computer,or video-game use alone or in combi-nation). In approximately one-half ofthe studies (14 of 29), television onlywas the primary outcome. Only 5 of theinterventions measured other screen-related behaviors, such as eatingwhile watching television or having atelevision in the bedroom (Table 2).

With regard to sample characteristics(see Table 2), the majority of the inter-ventions targeted children betweenthe ages of 5 and 11 years (20 of 29).When demographic information wasavailable, most study populations in-cluded 25% to 50% nonwhite partici-pants and 25% to 50% male children.Ten studies targeted high-risk chil-dren. Eight studies limited participa-tion to children who were overweightor obese, and 5 excluded children whodid not use a predetermined amount ofscreen time. Most studies (20 of 29)were conducted in the United States.

Design Characteristics

As shown in Table 3, the vast majorityof the studies included an intervention

and a comparison group. Five interven-tions used 1 group with a pretest-posttest design. Among studies with 2groups of participants, most used ran-domization to assign group member-ship. All studies reported baseline val-ues of the outcome variable, 5 studiesreported data collected during the in-tervention period, and only 2 studiesreported follow-up data. Four of thestudies included in the analyses re-ported adjusted postintervention data.Most (20 of 29) studies reported usinga valid data collection tool.

Heterogeneity and ModeratorAnalysesTable 4 contains the effect-size esti-mates calculated with a random-effects model for the 2 main groupingsof studies, those reporting data col-

lected during the intervention periodand those reporting data collected af-ter the intervention. The within-groupvariability was greater than would beexpected by chance, signifying possi-ble heterogeneity within each group.The between-groups Q statistic did notsupport the assumption of homogene-ity; no moderators were identified. Toassess themagnitude of heterogeneitypresent within each level of the poten-tial moderators, the I2 statistic wascalculated for each subgroup (ie,level) (Table 4). In general, I2 valueswere moderate to high, using gener-ally accepted I2 values of 25, 50, and 75(low, medium, and high, respec-tively).77,79 The presence of heterogene-ity within subgroups and the nonsig-nificance in between-groups �2 (Q)

TABLE 3 Study Design Characteristics of Studies Included in the Meta-analysis

Author (Year) No. ofStudyGroups

Random Allocationto Groups

Time Point of DataCollection in Relation to

Interventiona

Valid Data-Collection ToolUsed

During After Follow-up

Angelbuer91 (1998) 1 No X X NoChin et al44 (2008) 2 Yes X YesDennison et al42 (2004) 2 Yes X NoEisenmann et al45 (2008) 2 Yes X X YesEpstein et al43 (2008) 2 Yes X YesFaith et al112 (2001) 2 Yes X YesFord et al92 (2002) 2 Yes X YesFoster et al47 (2008) 2 Yes X YesGolan et al93 (1998) 2 Yes X YesGoldfield et al46 (2006) 2 Yes X YesGortmaker et al94 (1999) 2 Yes X YesGortmaker et al95 (1999) 2 No X YesHarrison et al96 (2006) 2 Yes X YesJason97 (1987) 1 No X NoJason et al98 (1993) 1 No X X NoJones et al113 (2008) 2 Yes X YesKipping et al109 (2008) 2 Yes X YesMauriello et al114 (2006) 2 Yes X NoMcCanna110 (1989) 1 No X X YesNemet et al99 (2005) 2 Yes X NoNiemeyer100 (1988) 2 Yes X NoNova et al101 (2001) 2 Yes X NoRobinson102 (1999) 2 Yes X YesRobinson et al103 (2003) 2 Yes X YesRobinson et al104 (2006) 2 Yes X YesSalmon et al105 (2008) 2 Yes X YesSege et al106 (1997) 1 No X NoSimon et al107 (2006) 2 Yes X YesWeintraub et al111 (2008) 2 Yes X X Yesa All studies collected data at baseline.


TABLE 4 Effect Sizes (95% Confidence Limits) and Between-Group Tests of Heterogeneity (Q), Assessment of the Amount of Variability Caused byRandom Error (I2), and Tests of Moderators for Studies Included in the Meta-analysis

Variable No. ofStudies

SMD (Random-Effects Model) Heterogeneity Within Subgroups Between-Groups�2 Test (Q) ofModerators

Within-Groups�2 Test (Q)

I2

PointEstimate

95% CLs Estimate,%

95% CLs

Lower Upper Lower, % Upper, %

SubgroupAll studies included 29Time point of data collectiona

During intervention 5 �1.904 �3.041 �0.767 13a 69.28 21.25 88.02After intervention 27 �0.148 �0.224 �0.071 50a 47.79 18.21 66.67Studies with postintervention data 27Type of dataAdjusted 4 �0.243 �0.401 �0.085 7a 57.88 0.00 86.00Raw 23 �0.112 �0.189 �0.035 32 31.65 0.00 58.91 2.123ComparisonsBefore vs after the intervention 5 �1.349 �2.601 �0.097 18a 77.52 45.74 90.69Treatment vs control 16 �0.101 �0.157 �0.044 12 0.00 0.00 48.68Treatment vs controladjusted data 4 �0.243 �0.401 �0.085 7 57.88 0.00 86.00Treatment 1 vs treatment 2 2 �0.032 �0.857 0.222 1 27.24 0.00 71.79 6.985OutcomeTelevision only 17 �0.108 �0.171 �0.045 16 0.00 0.00 0.00Total screen time 10 �0.211 �0.369 �0.054 31a 70.97 44.60 84.79 1.436Where study was foundLiterature search 23 �0.156 �0.238 �0.074 45a 51.11 21.09 69.71Reference list of another article 3 �0.17 �0.579 0.24 3 33.33 0.00 77.98 2.608Personal files 1 0.148 �0.213 0.509 0 0.00Publication typeJournal article 24 �0.135 �0.209 �0.062 41a 48.78 16.02 68.76Thesis/dissertation 3 �0.968 �2.204 0.269 10a 80.00 36.76 93.68 1.734Age group

�5 y 2 �0.105 �0.483 0.273 2 50.00 0.00 87.155–11 y 18 �0.125 �0.241 �0.008 28b 39.29 0.00 65.4112–18 y 6 �0.176 �0.304 �0.049 18a 72.22 35.86 87.97 0.501Not provided 1 �0.192 �0.449 0.064 0Nonwhite

�50% 4 �0.091 �0.211 0.029 3 0.00 0.00 0.00�50% 9 �0.256 �0.386 �0.126 13 38.46 0.00 71.68Unclear 14 �0.098 �0.199 0.004 23b 43.48 0.00 69.82 4.345MaleUnclear 2 �0.002 �0.587 0.584 2 50.00 0.00 87.15�50% 9 �0.225 �0.423 �0.027 23a 65.22 29.15 82.92�50% 16 �0.105 �0.162 �0.047 15 0.00 0.00 0.00 1.454CountryUnited States 19 �0.189 �0.294 �0.084 37a 51.35 17.70 71.24Non–United States 8 �0.074 �0.153 0.004 7 0.00 0.00 0.00 2.956OverweightNot specified 11 �0.212 �0.363 �0.06 19b 47.37 0.00 73.78�50% 4 0.002 �0.268 0.272 3 0.00 0.00 0.00�50% 12 �0.129 �0.229 �0.029 27a 59.26 23.02 78.44 1.963High-risk sampleNo 19 �0.142 �0.213 �0.071 32b 43.75 3.36 67.26Yes 8 �0.269 �0.665 0.126 20a 65.00 25.40 83.58 0.039Television-control deviceNo 19 �0.124 �0.199 �0.049 34a 47.06 9.61 68.99Yes 8 �0.3 �0.725 �0.035 15b 53.33 0.00 78.99 2.021Theory basedNo 11 �0.156 �0.349 0.037 16 37.50 0.00 69.26Yes 16 �0.145 �0.231 �0.059 35a 57.14 25.19 75.45 0.01Setting of interventionSchool 13 �0.15 �0.239 �0.061 26a 53.85 13.56 75.36Home 6 �0.933 �1.933 0.067 16a 68.75 26.29 86.75Other (includes unclear) 8 �0.109 �0.195 �0.024 7 0.00 0.00 0.00 2.891

REVIEW ARTICLES



test supported combining studies us-ing a random-effects model.

Overall Measure of Effect

When intervention effects were as-sessed after the intervention, moststudies showed a small favorable ef-fect of the intervention (Fig 2A). Individ-ual intervention effects ranged from�3.9891 to 0.466.111 Several interven-tions had large effects with large CIs,likely because of the small sample size.These did not greatly impact the over-all mean effect size. The overall meaneffect of all interventions included inthe analysis was small yet statisticallysignificant. After the intervention, theoverall SMD in means effect size was�0.148 (95% CI: �0.224 to �0.071)(Fig 2A) and Hedges g was �0.144(95% CI: �0.217 to �0.072) (data notshown).

During the intervention period (ie,while the intervention was being deliv-ered), the effect was large and statis-tically significant (SMD: �1.904 [95%CI: �3.041 to �0.767]) (Fig 2B). Like-

wise, Hedges g was �1.807 (95% CI:�3.069 to �0.545) (data not shown).Because of a lack of variance, 2 studies(identified with “a” in Fig 2) could notbe included in the analyses to calcu-late Hedges g. Because the SMD andHedges g differed only slightly and 2studies could not be included in calcu-lations of Hedges g because of a lack ofvariance, only the SMDs are presented.

Publication Bias

Visual inspection of the funnel plots(Figs 3 and 4) show potential publica-tion bias. On the basis of the Duval andTweedie trim-and-fill method, 4 studiesare missing (illustrated by filled cir-cles in Fig 3). If the missing studieswere included in the calculation of theoverall postintervention mean effectsize, the SMD (95% CI) would be�0.133 (�0.218 to �0.047), which isstill small but significant. The trim-and-fill method did not identify any missingstudies among the group of studiesrepresenting data during the interven-tion period (Fig 4). Assessment of pub-

lication bias with the failsafe N furthersupports the conclusion that the realeffect size is not 0. For the group ofstudies presenting postinterventiondata, an additional 255 studies report-ing no effect would have to be locatedand included in the analyses to nullifythe existing results. For the second setof studies, those presenting data col-lected during the intervention period,the failsafe N was 40. Sensitivity analy-ses did not identify any decisions orstudies that greatly impacted the over-all conclusions.

DISCUSSION

This meta-analysis shows that inter-ventions to reduce children’s screentime have a small but statistically sig-nificant effect after the interventionand a large statistically significant ef-fect during the intervention. Althoughstudies included in the meta-analysiswere heterogeneous, there was vari-ability greater than would be expectedby chance, no significant moderatorswere identified to explain this variabil-

TABLE 4 Continued

Variable No. ofStudies

SMD (Random-Effects Model) Heterogeneity Within Subgroups Between-Groups�2 Test (Q) ofModerators

Within-Groups�2 Test (Q)

I2

PointEstimate

95% CLs Estimate,%

95% CLs

Lower Upper Lower, % Upper, %

InterventionistResearcher 4 �0.384 �1.121 0.354 12a 75.00 30.64 90.99Other (includes unclear) 23 �0.138 �0.21 �0.066 39b 43.40 7.27 65.46 0.421Screen time is the primary outcomeYes 14 �0.134 �0.267 0 24b 45.83 0.00 70.95No 13 �0.156 �0.252 �0.06 25b 52.00 9.67 74.49 0.071Used valid data collection toolNo 9 �0.165 �0.393 0.064 21a 61.90 21.32 81.56Yes 18 �0.151 �0.23 �0.073 30b 43.33 1.17 67.51 0.011Source of the dataDocument 20 �0.2 �0.306 �0.095 38a 50.00 16.27 70.14Document and/or authors 7 �0.08 �0.154 �0.006 6 0.00 0.00 0.00 3.378Random allocation to groupsNo 6 �0.464 �0.974 0.047 18a 72.22 35.86 87.97Yes 21 �0.142 �0.213 �0.071 33b 39.39 0.00 64.05 1.496No. of study groups1 5 �1.349 �2.601 �0.097 18a 77.78 46.48 90.772 22 �0.146 �0.213 �0.079 33b 36.36 0.00 61.97 3.54

Three studies91,97,110 provided data during and after the intervention and are included in the calculation of the effect-size estimates both during and after the intervention.a P� .01.b P� .05.


ity. It is possible that because of thenumber of studies included in the anal-ysis, there was insufficient power todetect effectmoderators. It also is pos-sible that including studies with variedprimary outcomes impacted the abilityto detect moderators. The decision toinclude all interventions that targeted

screen time instead of interventionsthat only targeted screen time is wor-thy of further discussion. We hypothe-sized that the majority of interven-tions, especially those developedearlier, were developed to addresschildhood obesity using screen-timereduction as amechanism to decrease

weight status. Although our broad in-clusion criteria limit our ability to un-derstand specific mechanisms of howtelevision time uniquely impacts bodyfatness (ie, using an effect-modifiermodel), it increases the generalizabil-ity and clinical utility of the findings be-cause attempts to reduce body fatnessare almost always going to be multi-component. Thus, it was important tonot exclude studies that had ascreen-time reduction strategy sim-ply because other strategies alsowere used. This reflects our philo-sophical position that it is of greaterpublic health value to know that in-cluding a screen-time reductionstrategy is efficacious for reducingbody fatness than it is to know simplythat screen-time reduction works asa stand-alone strategy. These pointsnotwithstanding, the ability to testpotential moderators will improve asthe evidence base on screen-time in-terventions grows.

Several methods were used to in-crease confidence in the study results.On the basis of the results of the sen-sitivity analyses, the decision to calcu-late SMD instead of Hedges g did notimpact the conclusion. Although an at-tempt was made to identify unpub-lished works, none were identified. Vi-sual observation of funnel plots, theDuval and Tweedis trim-and-fillmethod, and the failsafe N were usedto assess the robustness of thisstudy’s conclusion. The funnel plot andthe trim-and-fill method identified fewpossible missing studies and the fail-safe N was large. Collectively, thesemethods support the validity of the pri-mary finding.

The Guide to Community PreventiveServices59 groups interventions intoseveral categories: interventions thatinclude provision of information only(interventions that try to changeknowledge, attitudes, or norms); be-havioral interventions (those that try

FavorsIntervention

FavorsComparison

Standard Mean Difference (SMD)

Study/First Author Point Estimate

UpperLimit

Standard Error

LowerLimitVariance

Angelbuer91 (1998) -3.983 -1.592 1.220 1.489 -6.375Chin44 771.0- 500.0070.0890.0930.0- )8002( Dennison42 408.0- 450.0232.0501.0053.0- )4002(

Eisenmann45 632.1- 232.0284.0256.0292.0- )8002(Epstein43 417.0- 060.0542.0842.0332.0- )8002( Ford92 827.0- 441.0973.0857.0510.0 )2002(

Foster47 902.0- 600.0770.0290.0850.0- )8002(

Golan93 711.1- 580.0292.0620.0645.0- )8991( Gortmaker94 284.0- 300.0850.0652.0-963.0- )9991( Gortmaker95 514.0- 210.0901.0410.0002.0- )9991( Harrison96 492.0- 310.0511.0651.0960.0- )6002(

*Jason97 873.5- 601.2154.1013.0435.2- )7891( *Jason98 242.5- 310.2914.1123.0064.2- )3991( Jones113 723.0- 700.0180.0800.0-761.0- )8002( Kipping109 282.0- 210.0111.0551.0360.0- )8002( Mauriello114 282.0- 700.0180.0530.0321.0- )6002( McCanna110 284.1- 922.0974.0693.0345.0- )9891( Nemet99 328.0- 880.0692.0933.0242.0- )5002( Niemeyer100 844.0- 710.0131.0460.0291.0- )8891(

Nova101 991.0- 330.0281.0415.0851.0 )1002( Robinson102 555.0- 120.0541.0510.0072.0- )9991( Robinson103 877.0- 860.0162.0342.0862.0- )3002(

Robinson 104 035.0- 220.0941.0550.0832.0- )6002( Salmon105 312.0- 430.0481.0905.0841.0 )8002( Sege106 802.0- 710.0031.0203.0740.0 )7991( Simon107 (2006) -0.139 -0.012 0.065 0.004 -0.266

Weintraub111 904.0- 002.0744.0243.1664.0 )8002(

-0.148 -0.071 0.039 0.002 -0.224

-4.00 -2.00 0.00 2.00 4.00

Mean Effect Size

A

Favors Intervention

Favors Comparison

tnioPStudy/First AuthorEstimate

UpperLimit

StandardError

Lower LimitVariance

Angelbuer91 (1998) -3.744 -1.479 1.155 1.335 -6.008 McCanna110 (1989) -0.402 0.509 0.465 0.216 -1.313 Goldfield46 (2006) -1.685 -0.851 0.426 0.181 -2.520 Faith112 (2001) -3.290 -1.405 0.962 0.925 -5.175 *Jason98 (1993) -1.658 0.478 1.090 1.187 -3.793

-1.904 -0.767 0.580 0.336 -3.041

-8.00 -4.00 0.00 4.00 8.00

Mean Effect Size

Standard Mean Difference (SMD)

B

FIGURE 2Forest plot of the mean overall SMD (95% CI) and the SMD and associated 95% confidence limits andthe associated SEs and variance for each study included in the analysis. The individual study effectsizes are represented by the center of the square symbol associated with the study, a study’s contri-bution toward the overall mean effect size (or weight) is represented by the symbol size, and theprecision of the study is represented by the length of the line associated with each symbol. A negativeeffect size denotes a program that favored the intervention group (eg, a significantly greater decreasein screen time in the intervention group compared with the treatment group). Generally acceptedcriteria for effect-size magnitude are 0.2� small, 0.5� medium, and 0.8� large.73,74

REVIEW ARTICLES



to change behavior by providing skillsor materials); environmental interven-tions (those that try to change thephysical and/or social environment);and policy or regulatory interventions.Most of the interventions reviewed in-clude an information provision compo-nent.42–45,47,91–107 This strategy has beenrecommended by Dennison and Ed-munds108 as a means to change behav-ior. The majority of the interventionsare behavioral interventions. They in-crease skills by having the child and/orparent develop a television-viewingbudget or plan, set screen-time goals

or monitor viewing,* or have the childidentify alternative activities.42,96,100,109

Several of the interventions attemptedto change behavior by modifying theenvironment by restricting access tothe television or computer using atelevision-control device† or by provid-ing opportunities for physical activ-ity.99,103,107,111 It is worth noting that491,97,98,110 of 5 interventions with largeeffect sizes after the intervention useda television-control device to help bud-

get screen time. Although the majorityof adolescents (aged 8–18 years) havemedia devices in their bedrooms,6 and33% of children aged 6 years andyounger have televisions in their bed-rooms,5 only 1 intervention42 reportedchanges in the proportion of studysubjects with televisions in the bed-room. Finally, none of the interventionsincluded in this review attempted tochange children’s behavior throughregulation. However, 1 intervention,107

included policy makers in the changeprocess by requesting that they pro-vide supportive environments forphysical activity in the form of low- orno-cost physical activity opportuni-ties. Given that most interventionsused a combination of strategies, in-tervention strategy was not tested asa moderator.

CONCLUSIONS

Children are able to take with themand use media now more than ever. Infact, 20% of children’s media con-sumption is from mobile devices, suchas cell phones, iPods/MP3 players, andvideo-game players.6 Reducing theamount of time children spend withscreen media and increasing discrim-inatemedia use andmedia use budget-ing are important given the negativehealth and behavioral implications ofexcessive screen time. Results fromthis meta-analysis show that interven-tions to decrease children’s screentime have a small but statistically sig-nificant effect. Because excessivescreen media use has been associatedwith many negative behavioral andhealth consequences, these resultssupport the implementation of screen-time reduction interventions. Many ofthe interventions reviewed can pro-vide children with the skills needed todecrease screen-media use. Parentsand clinicians should incorporate theinterventions or the strategies that arecommon to the effective interventionsinto their efforts to combat childhood

*Refs 44, 91, 95, 96, 98, 100, and 102–104.†Refs 43, 46, 91, 92, 97, 98, 102–104, and 110.

-4 -3 -2 -1 0 1 2 3 4

0.0

0.5

1.0

1.5

2.0

Stan

dard

err

or

Standard mean difference

FIGURE 3Funnel plot of the SE by SMD of studies reporting data from during and after the intervention. Observedstudies are indicated by open circles, and the imputed studies are indicated by filled circles.

-4 -3 -2 -1 0 1 2 3 4

0.0

0.5

1.0

1.5

2.0

Stan

dard

err

or

Standard mean difference

FIGURE 4Funnel plot of the SE by SMD of studies reporting data during the intervention. Observed studies areindicated by open circles. No missing studies were identified.


obesity. Even modest effects could re-sult in a positive change in the healthstatus of the population given the largenumber of children who use screenmedia and the increasing amount of

time children spend with media. As theevidence base expands, and the num-ber of screen-time interventions in-creases, future research can expandon these findings by examining the

clinical relevance and sustainability ofeffects observed, conducting a morethorough analysis of effect modifiers,and by identifying critical componentsof effective interventions.

REFERENCES

1. Krebs NF, Jacobson MS, the AmericanAcademy of Pediatrics Committee on Nu-trition. Prevention of pediatric overweightand obesity. Pediatrics. 2003;112(2):424–430

2. Council on Communications and Media.From the American Academy of Pediatrics:policy statement: media violence. Pediat-rics. 2009;124(5):1495–1503

3. Sisson SB, Church TS, Martin CK, et al. Pro-files of sedentary behavior in children andadolescents: the US National Health andNutrition Examination Survey, 2001–2006.Int J Pediatr Obes. 2009;4(4):353–359

4. Manganello JA, Taylor CA. Television expo-sure as a risk factor for aggressive behav-ior among 3-year-old children. Arch Pedi-atr Adolesc Med. 2009;163(11):1037–1045

5. Rideout V, Hamel E. TheMedia Family: Elec-tronic Media in the Lives of Infants, Tod-dlers, Preschoolers and Their Parents.Menlo Park, CA: Kaiser Family Foundation;2006. Report 7500

6. Rideout VJ, Foehr UG, Roberts DF. Genera-tion M2: Media in the Lives of 8- to 18-Year-Olds. Menlo Park, CA: Kaiser FamilyFoundation; 2010. Report 7500

7. Coon KA, Tucker KL. Television and chil-dren’s consumption patterns: a review ofthe literature. Minerva Pediatr. 2002;54(5):423–436

8. Dennison BA, Erb TA, Jenkins PL. Televisionviewing and television in bedroom associ-ated with overweight risk among low-income preschool children. Pediatrics.2002;109(6):1028–1035

9. Lumeng JC, Rahnama S, Appugliese D,Kaciroti N, Bradley RH. Television exposureand overweight risk in preschoolers. ArchPediatr Adolesc Med . 2006;160(4):417–422

10. Jordan AB, Robinson TN. Children, televi-sion viewing and weight status: summaryand recommendations from a CDC-sponsored expert panel meeting. Ann AmAcad Polit Soc Sci. 2008;615(1):119–132

11. Marshall SJ, Biddle SJH, Gorely T, CameronN, Murdey I. Relationships between mediause, body fatness and physical activity inchildren and youth: a meta-analysis. Int JObes. 2004;28(10):1238–1246

12. Thompson DA, Christakis DA. The associa-

tion between television viewing and irreg-ular sleep schedules among children lessthan 3 years of age. Pediatrics. 2005;116(4):851–856

13. Lowry R, Wechsler H, Galuska DA, Fulton JE,Kann L. Television viewing and its associa-tions with overweight, sedentary lifestyle,and insufficient consumption of fruits andvegetables among US high schoolstudents: differences by race, ethnicity,and gender. J Sch Health. 2002;72(10):413–421

14. Moriarty CM, Harrison K. Television expo-sure and disordered eating amongchildren: a longitudinal panel study. JCommun. 2008;58(2):361–381

15. Harrison K. Television viewing, fat stereo-typing, body shape standards, and eatingdisorder symptomatology in grade schoolchildren. Commun Res . 2000;27(5):617–640

16. Vandewater EA, Bickham DS, Lee JH. Timewell spent? Relating television use to chil-dren’s free-time activities. Pediatrics.2006;117(2). Available at: www.pediatrics.org/cgi/content/full/117/2/e181

17. Huesmann LR, Taylor LD. The role of mediaviolence in violent behavior. Annu Rev Pub-lic Health. 2006;27:393–415

18. Joshi PT, Parr AF, Efron LA. TV coverage oftragedies: what is the impact on children?Indian Pediatr. 2008;45(8):629–634

19. Dietz WH, Gortmaker SL. Do we fatten ourchildren at the television set? Obesity andtelevision viewing in children and adoles-cents. Pediatrics. 1985;75(5):807–812

20. Singer DG, Singer JL.Handbook of Childrenand the Media. Thousand Oaks, CA: Sage;2001

21. Calvert SL, Wilson BJ. The Handbook ofChildren, Media, and Development. Mal-den, MA: Blackwell; 2008

22. Preiss RW, Gayle BM, Burrell N, Allen M,Bryant J, eds. Mass Media EffectsResearch: Advances Through Meta-analysis. New York, NY: Lawrence ErlbaumAssociates; 2007

23. Wartella E, Robb M. Historical and recur-ring concerns about children’s use of themass media. In: Calvert SL, Wilson BJ, eds.The Handbook of Children, Media, and De-

velopment. Malden, MA: Blackwell; 2008:7–26

24. Hampl JS, Wharton CM, Taylor CA, WinhamDM, Block JL, Hall R. Primetime televisionimpacts on adolescents’ impressions ofbodyweight, sex appeal, and food and bev-erage consumption. Nutr Bull. 2004;29(2):92–98

25. Harrison K, Marske AL. Nutritional contentof foods advertised during the televisionprograms children watch most. Am J Pub-lic Health. 2005;95(9):1568–1574

26. Witt SD. The influence of television on chil-dren’s gender role socialization. ChildEduc. 2000;76(5):322–324

27. Barner MR. Sex-role stereotyping in FCC-mandated children’s educational televi-sions. J Broadcast Electron. 1999;43(4):551–564

28. WilsonBJ, SmithSL, PotterWJ, et al. Violence inchildren’s television programming: assessingthe risks. J Commun. 2002;52(1):5–35

29. Will KE, Porter BE, Geller ES, DePasqualeJP. Is television a health and safety haz-ard? A cross-sectional analysis of at-riskbehavior on primetime television. J ApplSoc Psychol. 2005;35(1):198–222

30. Greenberg BS, Eastin M, Hofschire L,Lachian K, Brownell KD. Portrayals of over-weight and obese individuals on commer-cial television. Am J Public Health. 2003;93(8):1342–1348

31. Collins RL, Elliott MN, Berry SH, et al. Watch-ing sex on television predicts adolescentinitiation of sexual behavior. Pediatrics.2004;114(3). Available at: www.pediatrics.org/cgi/content/full/114/3/e280

32. Bell RA, Berger CR, Cassady D, TownsendMS. Portrayals of food practices and exer-cise behavior in popular American films. JNutr Educ Behav. 2005;37(1):27–32

33. Byrd-Bredbenner C, Finckenor M, GrassoD. Health related content in prime-timetelevision programming. J Health Com-mun. 2003;8(4):329–341

34. Story M, Faulkner P. The prime time diet: acontent analysis of eating behavior andfood messages in television program con-tent and commercials. Am J Public Health.1990;80(6):738–740

35. Tirodkar MA, Jain A. Food messages on Af-

REVIEW ARTICLES


www.pediatrics.org/cgi/content/full/117/2/e181





rican American television shows. Am JPublic Health. 2003;93(3):439–441

36. Silver Wallace L, Leenders N. Content anal-ysis of prime-time television coverage ofphysical activity, 1970–2001. Am J PrevMed. 2004;26(2):130–134

37. Shrum LJ. Television and persuasion: ef-fects of the programs between the ads.Psychol Market. 1999;16(2):119–140

38. Lewis MK, Hill AJ. Food advertising on Brit-ish children’s television: a content analy-sis and experimental study with nine-yearolds. Int J Obes. 1998;22(3):206–214

39. Henderson VR, Kelly B. Food advertising inthe age of obesity: content analysis of foodadvertising on general market and AfricanAmerican television. J Nutr Educ Behav.2005;37:191–196

40. Kennedy C. Examining television as an in-fluence on children’s health behaviors. JPediatr Nurs. 2000;15(5):272–281

41. Rich M. Is television healthy? The medicalperspective. In: Pecora N, Murray JP, War-tella EA, eds. Children and Television: FiftyYears of Research.Mahwah, NJ: LawrenceErlbaum Associates; 2007:109–147

42. Dennison BA, Russo TJ, Burdick PA, Jen-kins PL. An intervention to reduce televi-sion viewing by preschool children. ArchPediatr AdolescMed. 2004;158(2):170–176

43. Epstein LH, Roemmich JN, Robinson JL, etal. A randomized trial of the effects of re-ducing television viewing and computeruse on body mass index in young children.Arch Pediatr Adolesc Med. 2008;162(3):239–245

44. Chin APMJ, Singh AS, Brug J, van MechelenW. Why did soft drink consumption de-crease but screen time not? Mediatingmechanisms in a school-based obesityprevention program. Int J Behav Nutr PhysAct. 2008;5:41

45. Eisenmann JC, Gentile DA, Welk GJ, et al.SWITCH: rationale, design, and implemen-tation of a community, school, and family-based intervention tomodify behaviors re-lated to childhood obesity. BMC PublicHealth. 2008;8:223

46. Goldfield GS, Mallory R, Parker T, et al. Ef-fects of open-loop feedback on physical ac-tivity and television viewing in overweightand obese children: a randomized, con-trolled trial. Pediatrics. 2006;118(1). Avail-able at: www.pediatrics.org/cgi/content/full/118/1/e157

47. Foster GD, Sherman S, Borradaile KE, et al.A policy-based school intervention to pre-vent overweight and obesity. Pediatrics.2008;121(4). Available at: www.pediatrics.org/cgi/content/full/121/4/e794

48. Marshall SJ, Gorely T, Biddle SJ. A descrip-tive epidemiology of screen-based mediause in youth: a review and critique. J Ado-lesc. 2006;29(3):333–349

49. American Academy of Pediatrics, Commit-tee on Public Education. Children, adoles-cents, and television. Pediatrics. 2001;107(2):423–426

50. The Cochrane Collaboration. The CochraneLibrary. Available at: www3.interscience.wiley.com/cgi-bin/mrwhome/106568753/home. Accessed December 6, 2008

51. University of York. Centre for Reviews andDissemination. Available at: www.crd.york.ac.uk/crdweb. Accessed December 6, 2008

52. Summerbell CD, Ashton V, Campbell KJ, Ed-munds L, Kelly S, Waters E. Interventionsfor treating obesity in children. CochraneDatabase Syst Rev. 2003;(3):CD001872

53. Oude Luttikhuis H, Baur L, Jansen H, et al.Interventions for treating obesity in chil-dren. Cochrane Database Syst Rev. 2009;(1):CD001872

54. Rosenthal R. The “file drawer problem”and tolerance for null results. PsycholBull. 1979;86(3):638–641

55. Littell JH, Corcoran J, Pillai V. SystematicReviews and Meta-analysis. New York, NY:Oxford University Press; 2008

56. Zaza S, Briss PA, Harris KW, eds. The Guideto Community Preventive Services. WhatWorks to Promote Health? Task Force onCommunity Preventive Services. New York,NY: Oxford University Press; 2005

57. Briss PA, Zaza S, Pappaioanou M, et al. De-veloping an evidence-based guide to com-munity preventive services: methods. TheTask Force on Community Preventive Ser-vices. Am J Prev Med. 2000;18(1 suppl):35–43

58. Briss PA, Dolan Mullen P, Hopkins DP.Methods used for reviewing evidence andlinking evidence to recommendations. In:Zaza S, Briss PA, Harris KW, eds. The Guideto Community Preventive Services: WhatWorks to Promote Health? Task Force onCommunity Preventive Services. New York,NY: Oxford University Press; 2005

59. Zaza S, Wright-De Aguero LK, Briss PA, et al.Data collection instrument and procedurefor systematic reviews in the Guide toCommunity Preventive Services. TaskForce on Community Preventive Services.Am J Prev Med. 2000;18(1 suppl):44–74

60. Briss PA, Portnoy B, Vogel-Taylor M, Zaza S.Continuing research needs. In: Zaza S,Briss PA, Harris KW, eds. The Guide to Com-munity Preventive Services: What Works toPromote Health? Task Force on Commu-

nity Preventive Services. New York, NY:Oxford University Press; 2005

61. Kahn EB, Ramsey LT, Heath GW, et al. In-creasing physical activity.MMWR RecommRep. 2001;50(RR-18):1–16

62. van Sluijs EMF, McMinn AM, Griffin SJ. Ef-fectiveness of interventions to promotephys ica l act iv i ty in chi ldren andadolescents: systematic review of con-trolled trials. BMJ. 2007;335(7622):703

63. Lipsey MW, Wilson DB. Practical Meta-analysis. Thousand Oaks, CA: Sage; 2001

64. Petticrew M, Roberts H. Systematic Re-views in the Social Sciences: A PracticalGuide. Malden, MA: Blackwell; 2006

65. Mullen PD, Ramirez G. The promise and pit-falls of systematic reviews. Annu Rev Pub-lic Health. 2006;27:81–102

66. Higgins JPT, Green SB, eds. CochraneHandbook for Systematic Reviews of Inter-ventions . 4 .2.6. Chichester, UnitedKingdom: John Wiley & Sons; 2006

67. Pigott TD. Methods for handling missingdata in research synthesis. In: Cooper H,Hedges LV, eds. The Handbook of ResearchSynthesis. New York, NY: Russell SageFoundation; 1994:163–175

68. Borenstein M, Hedges LV, Higgins JPT,Rothstein HR. Introduction to Meta-analysis. West Sussex, United Kingdom:John Wiley & Sons; 2009

69. Orwin RG. Evaluating coding decisions. In:Cooper H, Hedges LV, eds. The Handbook ofResearch Synthesis. New York, NY: RussellSage Foundation; 1994:139–162

70. Neuendorf KA. The Content Analysis Guide-book. Thousand Oaks, CA: Sage; 2002

71. Viera AJ, Garrett JM. Understanding inter-observer agreement: the kappa statistic.Fam Med. 2005;37(5):360–363

72. Borenstein M, Hedges L, Higgins J, Roth-stein H. Comprehensive Meta-analysis.2nd ed. Englewood, NJ: Biostat; 2005

73. Wolf FM.Meta-analysis: Quantitative Meth-ods for Research Synthesis. Vol 59, Illus-trated ed. Newbury Park, CA: Sage; 1986

74. Cohen J. Statistical Power Analysis for theBehavioral Sciences. Hillsdale, NJ: Law-rence Erlbaum Associates; 1988

75. Hedges LV. Statistical considerations. In:Cooper H, Hedges LV, eds. The Handbook ofResearch Synthesis. New York, NY: RussellSage Foundation; 1994:29–38

76. Higgins J, Thompson S, Deeks J, Altman D.Statistical heterogeneity in systematic re-views of clinical trials: a critical appraisalof guidelines and practice. J Health ServRes Policy. 2002;7(1):51–61

77. Huedo-Medina TB, Sánchez-Meca J, Marín-






www3.interscience.wiley.com/cgi-bin/mrwhome/106568753/home



www.crd.york.ac.uk/crdweb

www.crd.york.ac.uk/crdweb

Marínez F, Botella J. Assessing heteroge-neity in meta-analysis: Q statistic or I2 in-dex? Psychol Methods. 2006;11(2):193–206

78. Ioannidis JP, Patsopoulos NA, Evangelou E.Uncertainty in heterogeneity estimates inmeta-analyses. BMJ. 2007;335(7626):914–916

79. Higgins JPT, Thompson SG, Deeks JJ, Alt-man DG. Measuring inconsistency in meta-analyses. BMJ. 2003;327(7414):557

80. Higgins JPT, Thompson SG. Quantifyingheterogeneity in a meta-analysis. StatMed. 2002;21(11):1539–1558

81. Egger M, Davey Smith G, Schneider M,Minder C. Bias in meta-analysis detectedby a simple, graphical test. BMJ (Clin ResEd). 1997;315(7109):629–634

82. Becker BJ. Failsafe N or file-drawer num-ber. In: Rothstein HP, Sutton AJ, BorensteinM, eds. Publication Bias in Meta-analysis:Prevention, Assessment and Adjustments.Chichester, West Sussex, England: JohnWi-ley & Sons; 2005:111–125

83. Duval S, Tweedie R. Trim and fill: a simplefunnel-plot-based method of testing andadjusting for publication bias in meta-analysis. Biometrics. 2000;56(2):455–463

84. Peters JL, Sutton AJ, Jones DR, Abrams KR,Rushton L. Performance of the trim and fillmethod in the presence of publication biasand between-study heterogeneity. StatMed. 2007;26(25):4544–4562

85. Sterne JAC, Becker BJ, Egger M. The funnelplot. In: Rothstein HP, Sutton AJ, Boren-stein M, eds. Publication Bias in Meta-analysis: Prevention, Assessment and Ad-justments. Chichester, West Sussex,England: John Wiley & Sons; 2005:75–98

86. Duval S. The trim and fill method. In: Roth-stein HP, Sutton AJ, Borenstein M, eds.Publication Bias in Meta-analysis: Preven-tion, Assessment and Adjustments. Chich-ester, West Sussex, England: John Wiley &Sons; 2005:127–144

87. Johnson DB, Birkett D, Evens C, PickeringS. Statewide intervention to reduce televi-sion viewing in WIC clients and staff. Am JHealth Promot. 2005;19(6):418

88. Johnston BD, Huebner CE, Anderson ML,Tyll LT, Thompson RS. Healthy steps in anintegrated delivery system: child and par-ent outcomes at 30 months. Arch PediatrAdolesc Med. 2006;160(8):793

89. Epstein LH, Valoski AM, Vara LS, et al. Ef-fects of decreasing sedentary behaviorand increasing activity on weight changein obese children. Health Psychol. 1995;14(2):109–115

90. Epstein LH, Paluch RA, Gordy CC, Dorn J.Decreasing sedentary behaviors in treat-

ing pediatric obesity. Arch Pediatr AdolescMed. 2000;154(3):220–226

91. Angelbuer KC. Effects of Reducing Televi-sion Viewing on Six- to Eight-Year-Old Boys’Aggressive and Prosocial Behaviors [PhDdissertation]. Salt Lake City, UT: Universityof Utah; 1989

92. Ford BS, McDonald TE, Owens AS, RobinsonTN. Primary care interventions to reducetelevision viewing in African-Americanchildren. Am J Prev Med. 2002;22(2):106

93. Golan M, Fainaru M, Weizman A. Role ofbehaviourmodification in the treatment ofchildhood obesity with the parents as theexclusive agents of change. Int J Obes.1998;22(12):1217–1224

94. Gortmaker SL, Cheung LW, Peterson KE, etal. Impact of a school-based interdisciplin-ary intervention on diet and physical activ-ity among urban primary school children:eat well and keep moving. Arch PediatrAdolesc Med. 1999;153(9):975–983

95. Gortmaker SL, Peterson K, Wiecha J, et al.Reducing obesity via a school-based inter-disciplinary intervention among youth:Planet Health. Arch Pediatr Adolesc Med.1999;153(4):409–418

96. Harrison M, Burns CF, McGuinness M, Hes-lin J, Murphy NM. Influence of a health ed-ucation intervention on physical activityand screen time in primary schoolchildren: “Switch Off–Get Active.” J SciMed Sport. 2006;9(5):388–394

97. Jason LA. Reducing children’s excessivetelevision viewing and assessing second-ary changes. J Clin Child Psychol. 1987;16(3):245–250

98. Jason LA, Johnson SZ, Jurs A. Reducingchildren’s television viewing with an inex-pensive lock. Child Fam Behav Ther. 1993;15(3):45–54

99. Nemet D, Barkan S, Epstein Y, Friedland O,Kowen G, Eliakim A. Short- and long-termbeneficial effects of a combined dietary-behavioral-physical activity interventionfor the treatment of childhood obesity. Pe-diatrics. 2005;115(4). Available at: www.pediatrics.org/cgi/content/full/115/4/e443

100. Niemeyer K. Books and Beyond: ReadingAchievement, Reading for Pleasure, andTelevision Viewing Time [EdD disserta-tion]. Flagstaff, AZ: Northern ArizonaUniversity; 1987

101. Nova A, Russo A, Sala E. Long-termmanage-ment of obesity in paediatric officepractice: Experimental evaluation of twodifferent types of intervention. Ambul ChildHealth. 2001;7(3–4):239–247

102. Robinson TN. Reducing children’s televi-sion viewing to prevent obesity: a random-

ized controlled trial. JAMA. 1999;282(16):1561–1567

103. Robinson TN, Killen JD, Kraemer HC, et al.Dance and reducing television viewing toprevent weight gain in African-Americangirls: the Stanford GEMS pilot study. EthnDis. 2003;13(1 suppl 1):S65–S77

104. Robinson TN, Borzekowski DLG. Effects ofthe SMART Classroom Curriculum to Re-duce Child and Family Screen Time. J Com-mun. 2006;56(1):1–26

105. Salmon J, Ball K, Hume C, Booth M, Craw-ford D. Outcomes of a group-randomizedtrial to prevent excess weight gain, reducescreen behaviours and promote physicalactivity in 10-year-old children: switch-play. Int J Obes. 2008;32(4):601–612

106. Sege RD, Perry C, Stigol L, et al. Short-termeffectiveness of anticipatory guidance toreduce early childhood risks for subse-quent violence. Arch Pediatr Adolesc Med.1997;151(4):392–397

107. Simon C, Wagner A, Platat C, et al. ICAPS: amultilevel program to improve physicalactivity in adolescents. Diabetes Metab.2006;32(1):41–49

108. Dennison BA, Edmunds LS. The role of tele-vision in childhood obesity. Prog PediatrCardiol. 2008;25(2):191–197

109. Kipping RR, Payne C, Lawlor DA. Ran-domised controlled trial adapting USschool obesity prevention to England. ArchDis Child. 2008;93(6):469–473

110. McCanna MW. The Effects of a TelevisionWatching Reduction Program on Televi-sion Viewing and Alternative Activities[PhD dissertation]. Chicago, IL: DePaulUniversity; 1989

111. Weintraub DL, Tirumalai EC, Haydel KF, Fu-jimoto M, Fulton JE, Robinson TN. Teamsports for overweight children: the Stan-ford Sports to Prevent Obesity RandomizedTrial (SPORT). Arch Pediatr Adolesc Med.2008;162(3):232–237

112. Faith MS, Berman N, Heo M, et al. Effects ofcontingent television on physical activityand television viewing in obese children.Pediatrics. 2001;107(5):1043–1048

113. Jones D, Hoelscher DM, Kelder SH, Hergen-roeder A, Sharma SV. Increasing physicalactivity and decreasing sedentary activityin adolescent girls: the IncorporatingMore Physical Activity and Calcium inTeens (IMPACT) study. Int J Behav NutrPhys Act. 2008;5:42

114. Mauriello LM, Driskell MMH, Sherman KJ,Johnson SS, Prochaska JM, Prochaska JO.Acceptability of a school-based interven-tion for the prevention of adolescent obe-sity. J Sch Nurs. 2006;22(5):269–277

REVIEW ARTICLES



APPENDIX 1 Example of a Systematic Literature Search Strategy Used to Identify Articles for theMeta-analysis

Search ID Search Term Results

S1 Television 9534S2 Screen time 44S3 Program 50 400S4 Trial 690S5 Intervention 9307S6 Experiment 560S7 S1 or S2 9561S8 S3 or S4 or S5 or S6 60 133S9 S7 and S8 162S10 S9 limiters: published date 198601–200812 141

S indicates search number; 198601, January 1986; 200812, December 2008.


DOI: 10.1542/peds.2010-2353 originally published online June 27, 2011; 2011;128;e193Pediatrics

and Barbara A. DennisonDayna M. Maniccia, Kirsten K. Davison, Simon J. Marshall, Jennifer A. Manganello

ReductionA Meta-analysis of Interventions That Target Children's Screen Time for

ServicesUpdated Information &

http://pediatrics.aappublications.org/content/128/1/e193including high resolution figures, can be found at:

Referenceshttp://pediatrics.aappublications.org/content/128/1/e193#BIBLThis article cites 77 articles, 12 of which you can access for free at:

Subspecialty Collections

http://www.aappublications.org/cgi/collection/screen_time_subScreen Timehttp://www.aappublications.org/cgi/collection/media_subMediafollowing collection(s): This article, along with others on similar topics, appears in the

Permissions & Licensing

http://www.aappublications.org/site/misc/Permissions.xhtmlin its entirety can be found online at: Information about reproducing this article in parts (figures, tables) or

Reprintshttp://www.aappublications.org/site/misc/reprints.xhtmlInformation about ordering reprints can be found online:

by guest on October 4, 2021www.aappublications.org/newsDownloaded from

http://http://pediatrics.aappublications.org/content/128/1/e193

http://pediatrics.aappublications.org/content/128/1/e193#BIBL

http://www.aappublications.org/cgi/collection/media_sub

http://www.aappublications.org/cgi/collection/screen_time_sub

http://www.aappublications.org/site/misc/Permissions.xhtml

http://www.aappublications.org/site/misc/reprints.xhtml

DOI: 10.1542/peds.2010-2353 originally published online June 27, 2011; 2011;128;e193Pediatrics

and Barbara A. DennisonDayna M. Maniccia, Kirsten K. Davison, Simon J. Marshall, Jennifer A. Manganello

ReductionA Meta-analysis of Interventions That Target Children's Screen Time for

http://pediatrics.aappublications.org/content/128/1/e193located on the World Wide Web at:

The online version of this article, along with updated information and services, is

by the American Academy of Pediatrics. All rights reserved. Print ISSN: 1073-0397. the American Academy of Pediatrics, 345 Park Avenue, Itasca, Illinois, 60143. Copyright © 2011has been published continuously since 1948. Pediatrics is owned, published, and trademarked by Pediatrics is the official journal of the American Academy of Pediatrics. A monthly publication, it

by guest on October 4, 2021www.aappublications.org/newsDownloaded from

http://pediatrics.aappublications.org/content/128/1/e193

a meta-analysis of interventions that target children’s

Documents