Association between obesity and education level among the ...

The Nutrition and Health Survey in Taiwan (NAHSIT) reported that the prevalence of obesity among Taiwanese adults was 22.8% in 2013–2016, ... Skiptomaincontent Thankyouforvisitingnature.com.YouareusingabrowserversionwithlimitedsupportforCSS.Toobtain thebestexperience,werecommendyouuseamoreuptodatebrowser(orturnoffcompatibilitymodein InternetExplorer).Inthemeantime,toensurecontinuedsupport,wearedisplayingthesitewithoutstyles andJavaScript. Advertisement nature scientificreports articles article AssociationbetweenobesityandeducationlevelamongtheelderlyinTaipei,Taiwanbetween2013and2015:across-sectionalstudy DownloadPDF Subjects DiseasesRiskfactors AbstractTheinverseassociationbetweenobesityandeducationlevelhasbeendemonstratedinmanydevelopedcountries;however,fewstudieshaveinvestigatedobesityingeriatricpopulations.Thiscross-sectionalanalysisexploredtheassociationbetweengeriatricobesityandeducationlevel,alongwithotherdemographiccharacteristicsinTaipei,Taiwanbetween2013and2015.Taipeicitizens≥ 65 years(aborigines≥ 55 years)wererecruitedtoparticipateintheelderlyhealthexaminationprogramme.Logisticregressionwasappliedtoanalysetherelationshipbetweenobesity(definedasbodymassindex≥ 27 kg/m2inTaiwan)andeducationlevelamongmenandwomenaftercontrollingforage,race,incomestatus,andsmokingstatus.Atotalof28,092menand31,835womenwereincludedinthefinalanalysis.Comparedtothosewitheducationyears≥ 16,oldermenandwomenwitheducationyears≤ 12hadhigheroddsofbeingobese.Theoddsratiosincreaseasyearsofeducationdecrease,andthetrendismorepronouncedamongwomen.Aborigineshadmuchhigherchancesofbeingobeseamongmenandwomen,whiletherewerenodifferencesbyincomestatus.Theresultsclarifiedthefactorsrelatedtoobesityintheelderly,andwillbeusefulforauthoritiesworkingtoimprovehealthoutcomesamongthispopulation. DownloadPDF IntroductionTheUnitedNations(UN)andtheWorldHealthOrganization(WHO)defined‘ageing’,‘aged’,and‘super-aged’societiesaspopulationswheretheproportionofpeopleaged65 yearsandolderrangesfrom7to14%,> 14%,and> 20%,respectively.Manydevelopedcountriesaregraduallymovingtowardsagedandsuper-agedsocieties,includingTaiwan.Taiwanadvancedfromanageingsocietytoanagedsocietyinonly25 years(1993–2018),muchfasterthanothercountries1,2,3.InTaipei,especially,thepopulationaged65andovergrewmuchfasterthanthepopulationintherestofTaiwan;therefore,asearlyas2014,Taipeihadbecomeanagedsociety4.ItisestimatedthatTaiwanwillentertheeraofsuper-agedsocietyin2026,andtheTaiwangovernmentneedstoadaptquicklytothisnewreality1. Atthesametime,theworldwideprevalenceofobesityinadultshasnearlytripledsince1975;39%ofadults≥ 18 yearsoldwereoverweight,and13%wereobesein20165.TheNutritionandHealthSurveyinTaiwan(NAHSIT)reportedthattheprevalenceofobesityamongTaiwaneseadultswas22.8%in2013–2016,higherthanthatin1993–1996(11.5%)andin2005–2008(17.9%).Geriatricobesityhasalsoincreasedovertime.TheprevalenceofobesityamongolderadultsinTaiwanwas18.9%in1993–1996,22.2%in2005–2008,and22.8%in2013–20166.AllthesestatisticstestifytothegravityofgeriatricobesityasaforthcomingmajorhealthissueinTaiwan.Thecurrentevidenceshowsthat,whileoverweightmayserveasaprotectionagainsttheriskofdeathinthegeneralcommunity-dwellingelderlypopulation,obesityappearstobeassociatedwithastatisticallysignificantincreaseinriskofdeathamongtheelderly7.Obesityalsohaspositivecorrelationswithaplethoraofchronicdiseasesandmedicalproblems,includinghypertension,dyslipidemia,type2diabetes,coronaryheartdisease,andeventheincidenceofcancers8,9,10,11,resultinginsignificanthealthburdens.PreviousreportshaveshowndifferencesintheprevalenceofobesityamongadultsintheUnitedStatesbysex,agegroup,race,householdincome,educationlevel,smokingstatus,andurbanisationlevel12,13.Adultslivinginnonmetropolitanstatisticalareashadasignificantlyhigherprevalenceofobesityandsevereobesitycomparedwiththoselivinginlargemetropolitanstatisticalareas12.Bothwomenandmenwhowerecollegegraduateshadlowerprevalenceofobesitythanthosewithlesseducation12,13.InareportfromtheOrganizationforEconomicCo-operationandDevelopment(OECD),developedcountries,suchasAustralia,Canada,andEngland,showedsimilarresultstotheUSinthattheprevalenceofobesitywasloweramongthosewithhighereducationlevels14.DatafromKoreaNationalHealthandNutritionExaminationSurvey(KNHANES)in1998–2005and2010–2012disclosedthateducationwaspositivelyassociatedwithobesityinmen,i.e.,peoplewithhighereducationlevelsweremorelikelytobeobese,buttheresultsmightdifferinmodelswithandwithoutinteraction-effecttermsofindependentvariables.ThepatternsdifferedsubstantiallybetweensexesinKorea,andwomenwithhighereducationlevelshadasignificantlylowerpossibilityofbeingobese,thesameasotherdevelopedcountries14,15,16.AreportonhealthinequalitiesinTaiwanalsofoundthatdifferencesintheratesofoverweightandobesityamongwomenweresubstantial.Womenwithlessthanjuniorschooleducationhadafarhigherrateofoverweightorobesitythanothergroups,andgroupswithhighereducationlevelshadalowerprevalenceofoverweightorobesity17.AnotherTaiwanesestudyshowedmorbidobesitytobeassociatedwithlowsocioeconomicstatus(SES)18.Previousstudiesplacedlessemphasisonobesityamongelderlypopulations.OnestudycomparingtheelderlyinJapanandtheUSshowedthateachyearofeducationreducedtherelativelikelihoodofobesityby5–9%.Thisstudyhadsomelimitations,however,includinginsufficientnumbersofobesityamongJapaneseolderadultsandtheself-reporteddatausedintheanalysis19.Sincegeriatricobesitywillcauseanincreaseingovernmentexpenditureonhealthcareinthenearfuture,thereisanurgentneedforidentifyingtheassociatedriskfactors.EvidenceshowedmultiplepathwaysfromSES(includingeducationlevel,income,andoccupation)tohealth,includingdifferentialexposuretochronicstressanditsbiologicaltoll20.Establishingthecorrelationbetweengeriatricobesityandeducationlevelwillallowrelevantauthoritiestodevelopeffectiveinterventionstoreducetheprevalenceofobesity.Inthisstudy,weperformedacross-sectionalanalysistoexploretheassociationbetweengeriatricobesityandeducationlevel,alongwithotherdemographiccharacteristicsinTaipei,Taiwan.MaterialsandmethodsThiswasacross-sectionalstudyusingmultivariatelogisticregressionanalysistoevaluatetherelationshipbetweeneducationlevelandobesityamongtheelderly.DatawerecollectedfromtheelderlyhealthexaminationprogrammeinTaipeifrom2013to2015.TheelderlyhealthexaminationprogrammeisconductedannuallyinTaipeicityandfundedbythemunicipalgovernment.TherawdataweremadeavailableafterapplyingtotheDepartmentofHealth,TaipeiCityGovernment.Thisstudywasconductedbyanalysingdatasets,andtherawdatawerede-identified,thereforetheResearchEthicsCommitteeagreedtowaivetheinformedconsentduetominimalriskwithinthestudy.ThestudywasapprovedbytheTaipeiCityHospitalResearchEthicsCommitteewiththecasenumberTCHIRB-10805022-W.AllmethodswerecarriedoutinaccordancewithrelevantguidelinesandregulationsoftheTaipeiCityHospitalResearchEthicsCommittee.Taipeicitizensolderthan65 yearsoldarequalifiedtosignupfortheannualelderlyhealthexaminationatnocost.Asforthecitizenswithaboriginalidentity,criteriaofregistrationwereexpandedto55 yearsorolder.AllthehealthexaminationswereconductedatthecontractedhospitalsinTaipei.Theparticipantsunderwentsomeofthefollowingcheck-upitems:physicalexamination,bloodtest,urineanalysis,chestX-ray,electrocardiogram,andabdominalultrasonography.Well-trainedinterviewersalsoperformedtheAD8dementiascreeninganddepressionscreeningtest.Participants’backgroundinformation,suchaseducationlevel,smokingstatus,andpastmedicalhistory,wasobtainedusingaquestionnairedesignedbytheDepartmentofHealth,TaipeiCityGovernment.Heightandweightweremeasuredsimilarlyinallthecontractedhospitalsusingstandardisedtechniques,andparticipantswereaskedtoremovetheirshoesbeforethemeasurementsweretaken.Thedevicesusedinthemeasurementofbodyheightandweightwereallcalibratedregularlyunderthehospitals’standardoperatingprocedures.Obesityinthisstudywasdefinedasbodymassindex(BMI) ≥ 27 kg/m2,accordingtotheMinistryofHealthandWelfareofTaiwan.Thedefinitionsofobesity(BMI ≥ 30 kg/m2)andoverweight(BMI:25to 110 years,bodyweight> 120 kgor 200 cmor 50 kg/m2or 80 yearsold.AllanalyseswereconductedusingSASversion9.4(SASInstituteInc).ResultsTheinitialdatabaseincluded42,105,37,318,and37,544originaldatafrom2013,2014,and2015,respectively.Afterexcludingdatawithmissinginformationanderroneousvalues,dataforthesameparticipantsfromdifferentyears,andparticipantsmeetingtheexclusioncriteria,atotalof28,092maleparticipantswithameanageof76.8 yearsand31,835femaleparticipantswithameanageof74.2 yearswereincludedinthefinalanalysis.TheflowdiagraminFig. 1depictsdataprocessflow.Figure1Dataacquisitionflowdiagram.aForparticipantshavinghealthexaminationsinmorethanoneyear,onlythedataofthelatestyearwasincludedinfurtheranalysis.FullsizeimageDistributionofBMIinthesampleThesamplesizesforenrolledoldermenandwomenbyBMI,educationlevel,incomestatus,race,andsmokingstatusappearinTable1.Allcharacteristicsdifferedsignificantlybysex(pvalue  80 yearsoldamongmenandwomenenrolledfromtheelderlyhealthexaminationprogrammeinTaipei,2013–2015.FullsizetableDiscussionThisstudyrevealstheinverseassociationbetweenobesityandeducationlevelamongtheelderly,whichmeansthatobesityprevalenceandtheoddsofbeingobeseincreasewithdecreasingyearsofeducation.Comparedwiththosehavingmorethan16 yearsofeducation,theprevalenceandtheoddsratioshowedagradualbutnoticeableincrementasyearsofeducationdecreased.Thetrendwasmoreapparentamongwomen,andthegapsofprevalenceandoddsratiobetweeneachsubgroupofwomenweremoresignificantthanthatforsubgroupsofmen.AccordingtothedatareleasedfromtheDepartmentofBudget,AccountingandStatistics,TaipeiCityGovernment,thetotalpopulationofTaipeiover65 yearsoldwas399,182(men:181,035,women:218,147)in2015,andtheproportionofmenandwomenwithaneducationlevelbeloworequivalenttoprimaryschoolwas23.8%and44.8%,equivalenttojuniorhighschoolwas11.3%,and14.5%,equivalenttoseniorhighschoolwas21.2and19.8%,andequivalenttocollegeorabovewas43.8%and20.9%,respectively4.Therefore,theeducationalinequalitybetweenoldermenandwomendidexistinTaipei,asitdoesinTaiwan24,whichmightbeattributedtothesocio-culturalcontextandthetraditionalattitudestowardsgenderrolesatthetimewhentheywerestudying.AfterinspectingthedataofthisstudyandthedatafromTaipeiCityGovernmentindetail,itcanbesaidthatTaipeicitizenswithahighereducationlevelmightbemorelikelytoparticipateinelderlyhealthexaminations.Theactualstatisticsshowthat43.8%ofoldermenand20.9%ofolderwomeninTaipeihaveaneducationlevelequivalenttoorabovecollege,andinthisstudytheproportionofmaleandfemaleparticipantswitheducationyears≥ 13was48.2%and23.8%,respectively.Thelowereducationgrouppresentedtheoppositetrend,asthereare23.8%ofoldermenand44.8%ofolderwomeninTaipeihavinganeducationlevelbeloworequivalenttoprimaryschool,buttheproportionofmaleandfemaleparticipantswitheducationyears≤ 6inthisstudywasonly19.3%and36.2%,respectively.InTable3,theproportionofparticipantswitheducationyears≥ 16in2015wasgreaterthanin2013and2014,anditisknownthatrepeatedparticipantscomprisedmorethanhalftheparticipantsin2015,thereforesupportingtheaforementionedinference.Thismightindicateahigherlevelofhealthawarenessamongthemoreeducated.Accordingtopastindexstudiesandrecentliteraturethatdoesnotconsiderindividuals’ages,theresultsofdevelopingcountriesanddevelopedcountrieswithregardtoobesityprevalencehavebeendistinct.Apositiveassociationbetweenobesityandeducationlevelhasbeenmorecommonamongmeninlower-incomecountries,butaninverseassociationhasbeenmorecommonamongwomeninlower-incomecountriesandinbothsexesinhigher-incomecountries12,13,14,15,16,25,26,27,28.Taiwanisclassifiedasahigh-incomeeconomybytheWorldBankandasanadvancedeconomybytheInternationalMonetaryFund29,30.ThegovernmentofTaiwancalculateditsownHumanDevelopmentIndex(HDI)usingthecriteriaoftheUnitedNationsDevelopmentProgramme(UNDP),andthisHDI,havingavaluehigherthan0.800—thecut-offforthecategorisationintoaveryhighHDI—at0.911in2018,placedTaiwaninthecategoryofveryhighhumandevelopmentcountries31,32.Theresultsofthisstudyindicatedthattheassociationsmentionedabovewerenotchangedbythesubjects’agebeinglimitedtotheelderly.PreviousanalysesoftheUnitedStateshavealsoshownthattheprevalenceofobesityamongadultshasbeenlowerinthehigherincomegroup13,butsimilarresultswerenotobtainedinthisstudy,andthedifferencemightbeattributedtotheverylowproportionofparticipantswithalowerhouseholdincome(3.1%inmenand1.9%inwomen).Thepossibilityofbeingobesedecreasedwithageamongoldermenbutnotolderwomeninthisstudy,butthestudy’sresultsdidnotdisplayanycausalitybetweenageandBMIinbothsexes.Theresultofthiscross-sectionalanalysiswasinfluencedbytheagedistributionofalltheparticipantsandthepossiblecohorteffectwithindifferentagegroups.MuchhigherprevalenceofobesityamongtheyoungergroupsisnotedinTable4,andthisphenomenonmaybelinkedtothecohorteffectinrespecttoimprovednutritionalintake,ratherthanageitself.Despitethis,previouslongitudinalcohortstudiesrevealedthatBMI,aswellasbodyweightandheight,alldeclinewithageintheelderly.BMIdeclinesespeciallyaftertheageof70 years,butthetrendbecomesstatisticallyinsignificantaftertheageof90 years33,34.StudiesintheUnitedStateshaveshownthatracehasanapparenteffectonobesity3,12.However,theliteratureonTaiwaneseaboriginesonceshowedthataftercontrollingforothervariables,regressionanalysesrevealedfewassociationswithincreasedriskofobesityintheaborigines35.Thisstudyalsocontrolledforrelatedvariables,andstillfoundthatTaiwaneseaborigineshaveamuchhigherpossibilityofbeingobese.Pastliteratureonhealthdisparitiespointedoutthatevenifeducationlevelandincomewerecontrolled,therewerestillotherfactors,suchastheamountofwealthanddebt,thatcouldaffecthealthoutcomesamongdifferentracegroups20.Therefore,toclarifythecorrelationbetweenTaiwaneseaboriginesandobesity,moreprospectiveandwell-designedexperimentsarenecessary.Inthisstudy,oldersmokingmenhadlesspossibilityofbeingobese,butoldersmokingwomendidnothavethesametendency,whichmightbeduetothelowproportionofoldersmokingwomen.Occasionalsmokers,bothmenandwomen,showednodifferencesintermsofthepossibilityofbeingobesecomparedwithnon-smokers.Aprevioussystematicreviewshowedthatevenlightandintermittentsmokingcarriedhigherrisksoffrailtyandphysicaldisabilityinolderadults,andhigherrisksofcardiovasculardisease,respiratorydiseases,reproductivehealthconcerns,lungcancer,andgastrointestinalcancersinthegeneralpopulation.Thesamereviewalsoshowedthatcardiovascularmortalityandall-causemortalitywerepositivelyrelatedtooccasionalsmoking36.Thecorrelationbetweensmokingandbodyweightlossiswell-established,butthephysiologicalmechanismsarecomplexandincompletelyunderstood.Bodyweightisdeterminedbythebalanceofcaloricintakeandenergyexpenditure.Mostliteratureagreeswiththeeffectsofcigarettesmokingonbodyweightmediatedbynicotine,anditisgenerallyassumedthatnicotinereducesbodyweightbyincreasingenergyexpenditureandreducingappetite37,38.Itisworthnotingthatsmokingincreasesinsulinresistanceandisassociatedwithcentralfataccumulation,whichincreasestheriskofmetabolicsyndromeanddiabetes,resultinginanelevatedriskofcardiovasculardisease38.PreviousliteraturehasalsopointedoutthatcigarettesmokinghasbeentheleadingcauseofpreventablemorbidityandmortalityintheUnitedStates39,40.Therefore,thebenefitsofdecreasedbodyweightduetosmokingdonotovercometheassociativerisks.Thisisoneofthefewstudiesfocussingontherelationshipbetweenobesityandeducationlevelamongtheelderly,andthestudyalsohasseveralotherstrengths.Possibleconfoundersmentionedbypreviousstudies,includingage,race,smokingstatus,andincomestatus12,13,25,wereallcontrolledinthecurrentanalysis.Theresultswerestratifiedbysexbecausebackgroundinformationdifferedbetweenmenandwomen,andthiswouldhighlightanyexistingdifferences.Someearlierstudiesdidfinddifferencesrelatedtosex26,27,includingstudiesconductedinSouthKorea14,15,16,acountrywithasimilarculturalbackgroundtoTaiwan.Subgroupanalysesbydatayearandagegroupingwerealsotakenintoaccountinthisstudy.Therewereseverallimitationstothisstudy.Althoughthiscross-sectionalanalysisrevealedtheinverseassociationbetweenobesityandeducationlevelamongtheelderly,itwasunabletoelucidatethecausalrelationshipbetweenthetwo.Theuseofquasi-experimentaldesignsorfurtherlong-termprospectivecohortstudieswithindividualswithoutobesity,includingchildren,adolescents,adults,andtheelderly,mayassesstheimpactofeducationonBMImoreprecisely.Consideringthedatausedinthecurrentstudy,participantswithextremevaluesforage(> 110 years),bodyweight(> 120 kgor 200 cmor 50 kg/m2or