Assessment
Cross-cultural
adaptation
and
validation
of
the
health
literacy
assessment
tool
METER
in
the
Portuguese
adult
population
Dagmara
Paiva
a,b,c,
Susana
Silva
a,b,
Milton
Severo
a,b,
Pedro
Ferreira
d,
Osvaldo
Santos
e,
Nuno
Lunet
a,b,
Ana
Azevedo
a,b,*
a
DepartmentofClinicalEpidemiology,PredictiveMedicineandPublicHealth,UniversityofPortoMedicalSchool,Porto,Portugal
b
InstituteofPublicHealthoftheUniversityofPorto(ISPUP),Porto,Portugal
c
MonteMuradoHealthFamilyUnit,VilaNovadeGaia,Portugal
dInstituteofSocialSciences,UniversityofLisbon,Lisbon,Portugal e
InstituteofPreventiveMedicine,FacultyofMedicineofLisbon,Lisbon,Portugal
1. Introduction
Individualhealth literacy is thedegreetowhich individuals havethecapacitytoobtain,process,andunderstandbasichealth information and services needed to make appropriate health decisions[1].Inadequatehealthliteracyismoreprevalentamong theelderly,thosewithlowerlevelsofeducationalattainment[2] and withchronic disease[3]. Itis associated with poorer self-management skills,less successful navigationof thehealthcare system,highermorbidityandmortality[3–6].
TheEuropean HealthLiteracy Survey2011[7],conductedin eight European countries (Austria, Bulgaria, Germany, Greece, Ireland, the Netherlands, Poland, and Spain) found that only between 36.7% (in Spain) to 76.3% of the population (in the Netherlands) had adequate health literacy, as assessed by the Newest VitalSign [8].In Portugal, although health literacyhas startedtoappearinthenationalpoliticalagenda[9],thereareno publishedstudiesontheprevalenceofadequatehealthliteracy.
Healthliteracyiscommonlymeasuredusinginstrumentsbased onwordrecognitionorpronunciation:MedicalTermRecognition Test(METER)[10],RapidEstimateofAdultLiteracyinMedicine (REALM)[11],ShortAssessmentof HealthLiteracy for Spanish-speakingAdults(SAHLSA)[12],MedicalTerminologyAchievement Reading Test (MART) [13]; or reading comprehension and numeracy:NewestVitalSign(NVS)[8],TestofFunctionalHealth LiteracyinAdults(TOFHLA)[14].Mostinstrumentswereinitially developedinEnglishorSpanishandarebeingadaptedworldwide
ARTICLE INFO
Articlehistory:
Received27January2014
Receivedinrevisedform20May2014 Accepted13July2014 Keywords: Healthliteracy METER Validationstudies ABSTRACT
Objective:WeaimedtoculturallyadaptandvalidateMETERinthePortuguesepopulation,andtodefine
cut-offvaluesforadequatehealthliteracy.
Methods:We used the standardprocedure forthe adaptation of the words and surveyed health
professionalsto selectthenon-words.The instrumentwas administeredto atotalsample of249
participantsandretestedinasub-sampleof45afterthreemonths.Cut-offsweredefinedusingthe
modified Angoff procedure. Construct validity was assessed through association witheducational
attainmentandhealth-relatedoccupation.
Results:Exploratory factoranalysisrevealedtwodimensionsoftheinstrument,oneforwords and
anotherfornon-words.METERshowedahighdegreeofinternalconsistency,andacceptabletest–retest
reliability.Adequatehealthliteracywasdefinedasscoringatleast35/40inwordsand18/30in
non-words.Physiciansscoredhigherthananyothergroup,followedbyhealthresearchers,researchersfrom
otherareasandbypeoplewithprogressivelylowerlevelsofeducation(p<0.001).
Conclusion:Weculturallyadaptedabriefandsimpleinstrumentforhealthliteracyassessment,and
showeditwasvalidandreliable.
Practice implications: The Portuguese version of METER can be used to assess health literacy in
Portugueseadultsandtoexploreassociationswithhealthoutcomes.
ß2014ElsevierIrelandLtd.Allrightsreserved.
* Correspondingauthorat:DepartamentodeEpidemiologiaClı´nica,Medicina Preditiva e Sau´de Pu´blica,Faculdade de Medicina da Universidade do Porto, AlamedaProf.HernaˆniMonteiro,4200-319Porto,Portugal.Tel.:+351225513652; Fax:+351225513653.
E-mailaddress:[email protected](A.Azevedo).
ContentslistsavailableatScienceDirect
Patient
Education
and
Counseling
j ou rna l hom e pa ge :ww w. e l s e v i e r. c om/ l o ca t e / pa t e duc ou
http://dx.doi.org/10.1016/j.pec.2014.07.024
[15–19].Wordpronunciation-basedinstrumentsperformwellin Englishbut havefailed adaptationtolanguageswithvery high lettertosoundcorrespondence(suchasSpanish,Portugueseand Korean)becausetheyareunabletodiscriminatebetweenhealth literacyandabilitytoread[12,16,20].METERisbasedonword/ non-word recognition and is open-use, very brief, and self-administered, which means it can be added to a form or questionnairewithoutincreasingparticipantburdenconsiderably. We aimed to culturally adapt and validate METER in the Portuguesepopulation,andtodefinecut-offvaluesforadequate healthliteracy.
2. Methods
2.1. Originalinstrument
METERisanEnglishlanguageopenuseinstrumentbasedon REALM,consistingofalistof40medicalwordsand30made-up non-wordsthatintuitivelysoundlikerealmedicalterms.Itis self-administeredandittakesonaveragetwominutestocomplete.The participantsarerequestedtomarkonlythewordstheyaresureto beactualwords.Thescoreiscalculatedasthesumofallthecorrect wordsmarked. The original METER performance cut-off points definedbytheauthorswere0–20forlow,21–34formarginaland 35–40forfunctionalhealthliteracylevels[10].
2.2. Cross-culturaladaptation
Weusedthestandardprocedureforwordadaptation[21].An expertcommittee(withbackgroundsinFamilyMedicine,Internal Medicine,Pharmacy,Psychology,andSociology)ensured concep-tual and item equivalence. Afterwards, two native Portuguese speakersproficientlyfluentinEnglishtranslatedMETER indepen-dently and merged the translations into a single Portuguese version.Next, two native Englishspeakers, proficientin Portu-guese, independently back-translated this version. They were unawareofthepurposeoftheinstrumentandhadnotseenthe original version. The translators arrived at a consensus back-translatedversion,whichwasthenrevisedandcomparedtothe originalbythecommittee,resolvinganydiscrepanciesbetween thetwoversions.
For the non-words, we surveyed 25 health professionals to identifycommonmisspellingsandbuildupconstructionsbasedon realmedicalterms,30ofwhichwereselectedbytheresearchteam forinclusionintheinstrument.Theselectioncriteriaweretoavoid redundancyandtomaximizediversityofconceptualareas.
Thisversionwaspre-testedinasmallgroupofsixlaypeople andtheinstructionswordingwasadjustedforthesakeofclarity. 2.3. Sampleandrecruitment
The adapted version of the test was administered to a convenience sample of249 people from several heterogeneous groups:physicians(frompublichospitalsandprimarycarehealth centers),healthresearchers(fromaresearchinstitute),researchers fromareasunrelatedtohealth(fromanengineeringfaculty),and general population (from a primary care health center).In the absenceofprevalence dataofinadequatehealthliteracyin this population, the sample size was estimated based on other validation studies [8,11,12]. To assess construct validity we assumedthatphysicians wouldscorehigheston healthliteracy tests, followed by health researchers, people with a similar academicdegreeinareasunrelatedtohealth,andbypeoplewith progressivelylowerlevelsofeducationattainment.
Eligibilitycriteriafortheparticipantswereageover18years andabilitytospeakand readPortuguese.Potential participants
with impaired vision were excluded. The instrument was re-administeredtoaconveniencegroup(45healthresearchers)after athree-monthintervaltoassesstest–retestreliability.Thisrather longtest–retestintervaloftimeaimedtoreducemnesic/learning bias.
Thepresentinvestigationwascarriedoutinaccordancewith theCodeofEthicsoftheDeclarationofHelsinkiandapprovedby theEthics CommitteeofCentro HospitalardeSa˜o Joa˜oand the NationalCommitteeforDataProtection.Eachparticipantprovided writteninformedconsent.
2.4. Cut-offdefinition
Intheabsenceofagoldstandard,cut-offsweredefinedusing the modified Angoff procedure, a content-procedure method extensivelyappliedforestablishingabsoluteassessment criteria [22].Itisbased onexpertjudgmentofminimal competence of marginallycompetentindividuals:apanelofjudgestrainedinthe useofthemethoddiscussesandagreesonthecharacteristicsofa examineescoring‘‘borderline’’ for adequatehealth literacyand independentlyclassifieseachitemaccordingtothequestion‘‘Cana personwithminimalcompetenceanswertheitemcorrectly?’’,givena three-choiceoptionof‘‘yes’’,‘‘no’’or‘‘don’tknow’’.Anaverageofthe scoresofthejudgesiscalculatedtoprovideapassingscore(the cut-off).Thejudgesareallowedtoreviewanddiscusstheinitial scoresandaregiventheoptiontoindependentlyaltertheirown classificationiftheywishto;thisstrategyusuallydoesnotchange cut-offsmeaningfullybutreducesvariabilitybetweenjudges[23]. Thepanelcomprisedsixhealthliteracyexperts(withbackgrounds inFamilyMedicine,InternalMedicine,Pharmacy,Psychology,and Sociology).
Weusedthismethodtodichotomizehealthliteracylevelsinto adequateorinadequateinordertohelpguidedecisionstotailor patient education and communication interventions to the patients’needs,bothinfutureresearchandclinicalpractice. 2.5. Statisticalanalysis
Exploratoryfactoranalysiswasperformedonthe70itemsto evaluatehomogeneity(i.e.,toconfirmtherewasasinglelatent variablemeasuringwordrecognition)andCronbach’salphawas usedtomeasureinternalconsistency.Anitemwasconsideredto loadin adeterminedfactor whenit showedanabsolutefactor loadinghigherthan0.4.
Physicianswereexcludedfromtheseanalyses,sincetheyare not targetsof theinstrument.The globalgoodnessof fit ofthe underlyingmodelwasevaluatedusingthecomparativefitindex (CFI),recommendedforsamplesizesbelow250[24].
Logistic regression was used to compare the prevalence of adequatehealth literacyacross validation groups,adjusting for age.
Test–retestreliabilitywasassessedusingthestandarderrorof measurementandrespectivetwo-waymixedintra-class correla-tioncoefficientsingle-measure(ICC).
Exploratory factor analysis models were fitted using MPlus (V.5.2;Muthen&Muthen,LosAngeles,California,USA).Allother statisticalanalyseswereperformedusing Stataversion 11.1for Windows(StataCorpLP,CollegeStation,TX).
3. Results
Demographiccharacteristicsofthesamplebyvalidationgroup are summarized in Table 1. Women made up the majority of respondentsinallvalidationgroups(56.6%),exceptforthegroup ofresearchersinareasunrelatedtohealth(12.0%).Lesseducated peoplewereolder.
3.1. Exploratoryfactoranalysis
The scree plot curve inflected at the second component, revealingtwounderlyingdimensionsoftheinstrument(Fig.1). Exploratoryfactoranalysisconfirmedthesetwodimensions,with almostalltheitemsrepresentingrealwordsbeingincludedinthe firstdimension,andallthenon-wordsinthesecond(Table2).The CFIofthemodelimprovedfrom0.83intheuni-dimensionalmodel to0.93withthetwodimensions.
Fig.2depictstheplotofthepercentage ofcorrectlymarked wordsagainstthepercentageofnon-wordsthatwerecorrectlynot selected.Theformerrangedfrom15to100%,whereasthelatter from45to100%.Thesetwodimensionswerepoorlycorrelated (r=0.22)andthreepatternsemerged,visually:agroupofpeople scoringover two-thirds in both; a groupscoring lower in the identificationof realwords and a groupscoring lowerin non-words.Noone correctlyidentifiedless than two-thirdsof both wordsandnon-words.
3.2. Reliability
METERshowedahighdegreeofreliability,withaCronbach’s alphaof0.92forthefirstdimensionand0.83forthesecond.Inthe reteststudyafterthreemonths,thestandard errorof measure-mentwas1.54forthewords(ICC0.49)and0.82forthenon-words (ICC0.61).
3.3. Cut-offdefinition
Thefinalcut-offsdefinedbythejudgingpanelusingtheAngoff methodwere35 correctanswers inthewordssubscaleand 18 correctanswersinthenon-wordssubscale.Thereview,discussion andexperts’independentadjustmentoftheinitialscoreskept cut-offsroughlyunchanged(thecut-offfornon-wordsincreasedby onepointinthesecondroundandthecut-offforwordsremained
constant) and reduced the variability between judges (the standard deviationsdecreased from3.45to1.21andfrom2.71 to1.50,fornon-wordsandwords,respectively).
We used these cut-offs to categorize health literacy as inadequateoradequate;adequatewasdefinedasscoringatleast thecut-offvalueinbothwordsand non-words,i.e.35/40and 18/30,respectively.
3.4. Validity
Physicians scored higher than anyother group, followed by health researchers, non-health researchersand by people with progressivelylowerlevelsofeducation(Fig.3).Theage-adjusted prevalence of adequate health literacy increased consistently acrossvalidation groups (OR=2.79forphysicians, comparedto peoplewitheducationattainmentbelowthefourthgrade;pfor trend<0.001).
4. Discussionandconclusion 4.1. Discussion
Weculturallyadaptedahealthliteracyinstrumentthatisbrief andsimple,andshoweditwasvalidandreliable.Thisinstrument canbeusedtoassesshealthliteracylevelsandtosortbetween adequateandinadequatehealthliteracy.Weproposethatwords and non-words should be treated as different sub-scales with separatescoring.
Exploratory factor analysis revealed two dimensions of the instrument,oneforwordsandanotherfornon-words, implying thatsomeindividualsscoredhighinonedimensionandlowinthe other.Becauseofthedesignoftheoriginalinstrument,inwhich theparticipantsarerequiredtomarkonlythewordstheyaresure tobeactualwordsandnotindividuallymarkeachitemastrueor false,thenon-worditemslikelymeasuremorethanmerelyword recognition.Theresultssuggestthatthisdimensionalsomeasures riskaversion,i.e.individualsmoreaversetoerrormarklesswords and non-words, thus scoringlowerin thewordssub-scale and higher in the non-word subscale; the opposite applies to less cautiousindividuals.Thefactthatnoonescoredverylowinboth dimensionssupportsthisinterpretation.Withtheoriginalscoring instructions (total score as the sum of all the correct words marked),intheunlikelyeventthataparticipantmarkedallofthe items,he/shewouldachievethemaximumscore,eventhoughthat would not correspond to adequate health literacy. Therefore, scoringthetwodimensions independentlyand thencombining theperformanceonbothsub-scaleswillreducethe misclassifica-tionofindividualswithinadequatehealthliteracy.
The rare exceptions to theperfect correspondencebetween wordsandnon-wordsandeachofthetwodimensionswereitems 4(awordthatcouldfitineitherdimension),item37(anon-word thatcouldfitineitherdimension),anditem47(awordthatdoes not fit in either dimension). Only one individual neglected to correctlymarkitem4asaword,andthismighthavemisestimated the correlationdue toceiling effects, loadingthe item in both dimensions.Item37isoneofthegoodexamplesoftheeffectofthe small but significant correlation between the first and second dimensions-severalitemspartiallycross-loadedinboth(ascanbe observedbythesmallfactorloadingdifferencesbetweenfactors oneandtwo).Item47is‘‘impetigo’’,acontagiousskininfection whichcausessoresand blisters,and relativelyunfamiliartolay people.Veryfewnon-physicianscorrectlyselectedthiswordand thiscouldexplainwhyitdoesnotfitineitherdimension;itseems tobealmostexclusivelyrecognizedbyhealthprofessionals.Our finding is consistent with the word frequency effect in word 0 5 10 15 20 25 30 Eigenvalue
Component number
Fig.1.Screeplotofeigenvaluesafterexploratoryfactoranalysis. Table1
Demographiccharacteristicsofthesamplebyvalidationgroup. Validationgroup n Ageinyears,
mean(SD) Women,n(%) Physicians 53 32.3(7.1) 34(64.2) Healthresearchers 45 29.6(5.6) 37(82.2) Otherresearchers 50 43.8(13.0) 6(12.0) Generalpopulation Collegeeducation 18 41.6(13.7) 11(61.1) 12thgrade 15 34.8(11.1) 8(53.3) 9th–11thgrade 22 38.5(12.4) 14(63.6) 5th–8thgrade 17 41.4(14.0) 10(58.8) 4thgrade 29 61.1(9.2) 21(72.4) SD—standarddeviation.
Table2
Correctanswersperiteminwordsandnon-words,andstandardizedfactorloadingsforoneandtwofactorsinexploratoryfactoranalysis. Correctanswersn(%) Standardizedfactorloadings
Onefactor Twofactors
Factor1 Factor2 Words Portuguese English
METER2 Artrite Arthritis 173(88.3) 0.729 0.847 0.104
METER3 Obesidade Obesity 183(93.4) 0.849 0.952 0.105
METER4 Gripe Flu 195(99.5) 0.893 0.424 0.758
METER6 Sı´filis Syphilis 148(75.5) 0.668 0.885 0.280
METER7 Pota´ssio Potassium 161(82.1) 0.746 0.861 0.108
METER8 Hormonas Hormones 178(90.8) 0.745 0.819 0.015
METER9 Nervos Nerves 191(97.4) 0.647 0.600 0.224
METER14 Exercı´cio Exercise 176(89.8) 0.650 0.745 0.015
METER15 Pu´stula Pustule 63(32.1) 0.187 0.587 0.450
METER17 Rim Kidney 192(98.0) 0.882 0.890 0.142
METER18 Urgeˆncia Emergency 185(94.4) 0.617 0.633 0.115
METER20 Menopausa Menopause 192(98.0) 0.769 0.794 0.146
METER21 Diagno´stico Diagnosis 182(92.9) 0.906 0.947 0.019
METER23 Icterı´cia Jaundice 110(56.1) 0.466 0.769 0.333
METER24 Bexiga Bladder* 191(97.4) 0.907 0.900 0.131
METER25 Aborto Miscarriage 192(98.0) 0.888 0.870 0.175
METER26 Hepatite Hepatitis 189(96.4) 0.927 0.945 0.047
METER29 Asma Asthma 192(98.0) 0.898 0.897 0.151
METER30 Inflamato´rio Inflammatory 185(94.4) 0.734 0.836 0.019
METER31 Anemia Anemia 190(96.9) 0.880 0.923 0.054
METER34 Stress Stress 186(94.9) 0.848 0.906 0.003
METER39 Cancro Cancer 192(98.0) 0.902 0.878 0.145
METER41 Antibio´ticos Antibiotics 191(97.4) 0.870 0.893 0.104
METER43 Colite Colitis 116(59.2) 0.225 0.542 0.306
METER44 Diabetes Diabetes 194(99.0) 0.896 0.754 0.394
METER47 Impetigo Impetigo 10(5.1) 0.306 0.059 0.561
METER48 Menstrual Menstrual 186(94.9) 0.710 0.797 0.014
METER50 Convulsa˜o Seizure 175(89.3) 0.833 0.882 0.007
METER51 Apeˆndice Appendix 184(93.9) 0.866 0.938 0.058
METER54 Dose Dose 166(84.7) 0.754 0.895 0.187
METER55 Hemorro´idas Hemorrhoids 161(82.1) 0.303 0.396 0.017
METER56 Testı´culo Testicle 186(94.9) 0.931 0.982 0.065
METER57 Olho Eye 192(98.0) 0.851 0.850 0.174
METER61 Sexualmente Sexually 179(91.3) 0.596 0.674 0.018
METER64 Medicac¸a˜o Medication* 191(97.4) 0.731 0.759 0.131
METER65 Micro´bios Germs 174(88.8) 0.527 0.559 0.079
METER66 Gonorreia Gonorrhea 120(61.2) 0.529 0.820 0.342
METER68 Fadiga Fatigue 185(94.4) 0.778 0.841 0.020
METER69 Osteoporose Osteoporosis 186(94.9) 0.642 0.670 0.102 METER70 Obstipac¸a˜o Constipation 146(74.4) 0.518 0.742 0.184 Non-words
METER1 Imı´gdala N/A 176(89.8) 0.345 0.120 0.346
METER5 Nervosite N/A 173(88.3) 0.665 0.312 0.551
METER10 Anquia N/A 195(99.5) 0.868 0.426 0.759
METER11 Ca´stula N/A 195(99.5) 0.889 0.426 0.757
METER12 Ingesto N/A 173(88.3) 0.345 0.081 0.581
METER13 Intestigo N/A 181(92.3) 0.335 0.023 0.520
METER16 Cerpes N/A 194(99.0) 0.650 0.381 0.565
METER19 Xirope N/A 187(95.4) 0.623 0.370 0.458
METER22 Candı´ase N/A 167(85.2) 0.122 0.298 0.540
METER27 Enatoma N/A 185(94.4) 0.575 0.223 0.552
METER28 Unhal N/A 190(96.9) 0.455 0.031 0.723
METER32 Linsoma N/A 182(92.9) 0.328 0.022 0.509
METER33 Ceresiana N/A 158(80.6) 0.336 0.014 0.472
METER35 Alge´rico N/A 173(88.3) 0.295 0.067 0.337
METER36 Jezum N/A 166(84.7) 0.393 0.174 0.345
METER37 Su´rgico N/A 191(97.4) 0.643 0.440 0.378
METER38 Malorias N/A 195(99.5) 0.868 0.426 0.759
METER40 Alcoolidade N/A 160(81.6) 0.294 0.107 0.510
METER42 Antideprimido N/A 170(86.7) 0.395 0.111 0.640
METER45 Otorringologista N/A 138(70.4) 0.221 0.096 0.374
METER46 No´sea N/A 189(96.4) 0.491 0.116 0.635
METER49 Gatarral N/A 192(98.0) 0.444 0.138 0.572
METER52 Abdomina´vel N/A 151(77.0) 0.352 0.129 0.579
METER53 Enxuteca N/A 192(98.0) 0.589 0.244 0.622
METER58 Obste´rico N/A 183(93.4) 0.382 0.205 0.812
METER59 Sonambulac¸a˜o N/A 168(85.7) 0.321 0.211 0.664
recognition, in which low frequency words are less often recognizedaswords[25].
The instrument had a high internal consistency(Cronbach’s alpha0.80) in both domains, similar to that of the original instrument (Cronbach’s alpha=0.93) [10] and of other health literacytests[13,14].
Thetest–retestreliabilitywasonlyreasonablyacceptableasit iscontext-specificanddependsonhowmuchparticipantsdiffer fromeachother.Evenifthevariabilitybetweentheresultsinthe twotrialsisnegligible,theICCwillbesmalliftheretestgroupis homogenous [26], as was the case in the group of health researchers,whichisalimitationofthisstudy.
Thecut-offforthenon-wordscouldbeunderestimatedbythe Angoffmethodbecausethewayit is done doesnot reflect the instrument’sinstructions,previouslymentioned.Itmayhavebeen hardforthejudgestokeepcomingbacktotheconceptthatthetest scoredefaultfornon-wordsisinactionandthatthiscorrespondsto maximumscore.Answering‘‘yes/no/don’tknow’’tothequestion ofwhetheraminimallycompetentindividualwouldanswerthe question correctly is more suitable to items with a true/false format. Despite the good performance of the final version of theinstrument in discriminating the several validation groups, further studies comparing the performance of the non-word subscalewiththatofotherhealthliteracytestsareneeded.
Different health literacy assessment instruments categorize health literacy scores into a variable number of categories in addition to providing a continuous score [27]. We decided to dichotomizethescoresintoadequateandinadequateinsteadof maintaining the three categories of the original instrument to simplifythe decision-making regarding health education inter-ventions for patients with inadequate health literacy, both in researchandclinicalsettings.
Some validation studies of health literacy instrumentshave usedconcurrentvalidation,thatis,throughthecomparisonwith
an existinginstrument. Thisis a controversialoption giventhe multipleproposeddefinitionsoftheunderlyingconstruct[28]and thediverseandrestrictivescopeoftheinstruments[27].Thereis justnowaytotellwhichonebetterrepresentshealthliteracy.Our strategy assumed that health literacy should be higher in physicians,followedbyhealthresearchers,peoplewithasimilar academicdegreeinareasnon-relatedtohealthandbypeoplewith progressively lower levels of education attainment. The data confirm this hypothesis and this suggests that the instrument measures more than educational attainment, but we cannot excludethepossibilityofitnotmeasuringmorethantheabilityto recognize medical jargon—only one of the aspects of health literacy.Furthermore,METERandotherwordrecognitiontestsdo not directly address the individual ability of accessing, under-standing, processing and communicating information that is includedinthehealth literacyconstruct;vocabularyknowledge playsonlyasmallpartinthesecompetencies.However,thescore in theseinstrumentsis associatedwithotherclinicallyrelevant healthmeasuresandmaybeusedtoscreenforindividualswho could use more help in understanding and acting on health information.Thecomparisonwiththeperformanceofotherhealth literacy instruments may shed some light on this issue by exploring theneed tousemultipleinstrumentssimultaneously toassesshealthliteracy.
4.2. Conclusion
WeculturallyadaptedandvalidatedMETERinthePortuguese populationanddefinedcut-offvaluesforadequatehealthliteracy. This instrument distinctly differentiates individuals based on educationalattainmentandhealth-relatedoccupation,inspiteof measuringonlyvocabularyknowledge—asmallpartofthehealth literacyconstruct.Futurestudiesshouldrevealhowit performs whenusedtogetherwithhealthliteracyinstrumentsnotbasedon wordrecognition.
Table2(Continued)
Correctanswersn(%) Standardizedfactorloadings Onefactor Twofactors
Factor1 Factor2
METER62 Purisia N/A 193(98.5) 0.594 0.161 0.762
METER63 Fibro´mico N/A 184(93.9) 0.421 0.076 0.707
METER67 Esto´mico N/A 185(94.4) 0.571 0.117 0.700
Cronbach’salpha 0.894 0.916 0.828
N/A:Englishnotapplicableforthenon-words.
*
LiteraltranslationtoEnglishinsteadoftheoriginalversionwhenPortuguesewordswereadjustedtomaintainsemanticand/orstructuralequivalence.
Fig.2.Percentcorrectanswersinwordsandnon-wordsinMETER.
100 93 90 78 47 41 24 14 0 7 10 22 53 59 76 86 0% 20% 40% 60% 80% 100% Inadequate Adequate
4.3. Practiceimplications
ThePortugueseversionofMETERcanbeusedtoassesshealth literacy in Portuguese adults and to explore associations with health outcomes. Furtherstudies are neededto determine the usefulnessofthisinstrumentasascreeningtoolanddecision-aid inclinicalsettings,eitherusedonitsownorincombinationwith otherhealthliteracyassessmenttests.
Measuring the health literacy of Portuguese adults can highlighttheissueofinadequatehealth literacyinthenational politicalagenda,andraiseawarenessbythegeneralpopulation.In turn,thiscouldindirectlypromotesystemchangestoimprovethe communication of health information, namely by encouraging strategies that enhance comprehension by health consumers. Thesestrategieswillbenefitnotonlythosewithinadequatehealth literacybutpotentiallyeverybody.
Acknowledgments
WethankTeresaOliveira,GabrielCoutinho,JoePereiraandPaul Charlesforthetranslationandback-translationoftheinstrument. We are also grateful to each of the participants and to the institutionsFaculdadedeEngenharia daUniversidade do Porto, InstitutodeSau´dePu´blicadaUniversidadedoPortoandUnidade de Sau´de Familiar Monte Murado, for enabling participant recruitment.
This work was supported by Fundac¸a˜o para a Cieˆncia e a Tecnologia(HMSP-IISE/SAU-ICT/0004/2009).
AppendixA
Validatedversionoftheinstrument
Alistaseguinteincluialgunstermosqueexistemnalinguagem me´dica. Alguns desses termos esta˜o relacionados com partes ou func¸o˜esdocorpo,comtiposdedoenc¸asoucomcoisasquepodem melhoraroupiorarasau´de.Alistatambe´mconte´malgumaspalavras quepodemparecerousoarcomotermosreais,masquena˜oexistem. A` medidaqueforlendoestalista,coloqueumacruz‘‘X’’aoladodas palavrasquesa˜otermosreais.Na˜otenteadivinhar.Coloqueumacruz ‘‘X’’ ao lado daspalavras so´ quandotiver a certeza que existem mesmo. _______Imı´gdala _______Jezum _______Artrite _______Su´rgico _______Obesidade _______Malorias _______Gripe _______Cancro _______Nervosite _______Alcoolidade _______Sı´filis _______Antibio´ticos _______Pota´ssio _______Antideprimido _______Hormonas _______Colite _______Nervos _______Diabetes _______Anquia _______Otorringologista _______Ca´stula _______No´sea _______Ingesto _______Impetigo _______Intestigo _______Menstrual _______Exercı´cio _______Gatarral _______Pu´stula _______Convulsa˜o _______Cerpes _______Apeˆndice _______Rim _______Abdomina´vel _______Urgeˆncia _______Enxuteca _______Xirope _______Dose _______Menopausa _______Hemorroidas _______Diagno´stico _______Testı´culo _______Candı´ase _______Olho _______Icterı´cia _______Obste´rico _______Bexiga _______Sonambulac¸a˜o _______Aborto _______Drenac¸a˜o _______Hepatite _______Sexualmente _______Enatoma _______Purisia _______Unhal _______Fibro´mico _______Asma _______Medicac¸a˜o _______Inflamato´rio _______Micro´bios _______Anemia _______Gonorreia _______Linsoma _______Esto´mico _______Ceresiana _______Fadiga _______Stress _______Osteoporose _______Alge´rico _______Obstipac¸a˜o
Correctanswersinboldface. References
[1]Nielsen-BohlmanL,PanzerAM,KindigDA,editors.Healthliteracy:a prescrip-tiontoendconfusion.Washington,D.C.:TheNationalAcademiesPress;2004.
[2]KutnerM,GreenbergE,JinY,PaulsenC.ThehealthliteracyofAmerica’sadults: resultsfromthe2003NationalAssessmentofAdultLiteracy(NCES2006-483). Washington,DC:NationalCenterforEducationStatistics;2006.
[3]DeWaltDA,BerkmanND,SheridanS,LohrKN,PignoneMP.Literacyandhealth outcomes. J Gen Intern Med 2004;19:1228–39.Pubmed Central PMCID: PMC1492599.DOI:10.1111/j.1525-1497.2004.40153.x.
[4]AdHocCommitteeonHealthLiteracyfortheCouncilonScientificAffairs. Healthliteracy:reportoftheCouncilonScientificAffairs.AmericanMedical Association.JAMA1999;281:552–7.PubMedPMID:10022112.
[5]HerndonJB,ChaneyM,CardenD.Healthliteracyandemergencydepartment outcomes:asystematicreview.AnnEmergMed2011;57:334–45.PubMed PMID:21035902.DOI:10.1016/j.annemergmed.2010.08.035.
[6]SudoreRL,YaffeK,SatterfieldS,HarrisTB,MehtaKM,SimonsickEM,Newman AB,RosanoC,RooksR,RubinSM,AyonayonHN,SchillingerD.Limitedliteracy andmortalityintheelderly:thehealth,aging,andbodycompositionstudy.J Gen Intern Med 2006;21:806–12. http://dx.doi.org/10.1111/j. 1525-1497. 2006.00539.x.
[7]HLS-EUConsortium.ComparativereportofhealthliteracyineightEUmember states.In:TheEuropeanHealthLiteracySurveyHLS-EU;2012[cited2013Sep 30].Availablefromhhttp://www.healthliteracy.ie/wp-content/uploads/2012/ 09/HLS-EU_report_Final_April_2012.pdfi.
[8]WeissBD,MaysMZ,MartzM,etal.Quickassessmentofliteracyinprimary care:thenewestvitalsign.AnnFamMed2005;3:514–22.PubMedPMID: 16338915.DOI:10.1370/afm.405.
[9]Planonacionaldesau´de.2012-2016:3.1Eixoestrate´gico-cidadaniaemsau´de, [Nationalhealthplan2012-2016:3.1Strategicaxis-healthcitizenship],2012, [cited2013Sep30].Availablefrom:(http://pns.dgs.pt/pns-em-portugues/). [10]RawsonKA,GunstadJ,HughesJ,SpitznagelMB,PotterV,WaechterD,Rosneck
J.TheMETER:abrief,self-administeredmeasureofhealthliteracy.JGenIntern Med2009;25:67–71.http://dx.doi.org/10.1007/s11606-009-1158-7. [11]DavisTC,CrouchMA,LongSW,JacksonRH,BatesP,GeorgeRB,BairnsfatherLE.
Rapidassessmentofliteracylevelsofadultprimarycarepatients.FamMed 1991;23:433–5.PubMedPMID:1936717.
[12]LeeS-YD,BenderDE,RuizRE,ChoYI.Developmentofaneasy-to-useSpanish healthliteracytest.HealthServRes2006;41:1392–412.http://dx.doi.org/ 10.1111/j.1475-6773.2006.00532.x.
[13]Hanson-DiversEC.Developingamedicalachievementreadingtesttoevaluate patientliteracyskills:apreliminarystudy.JHealthCarePoorUnderserved 1997;8:56–69.PubMedPMID:9019026.
[14]ParkerRM,BakerDW,WilliamsMV,NurssJR.Thetestoffunctionalhealth literacyinadults:anewinstrumentformeasuringpatients’literacyskills.J GenInternMed1995;10:537–41.PubMedPMID:8576769.
[15]ConnorM,MantwillS,SchulzPJ.FunctionalhealthliteracyinSwitzerland— validationofaGerman,Italian,andFrenchhealthliteracytest.PatientEduc Couns2013;90:12–7. PubMedPMID:23089240.DOI:10.1016/j.pec.2012. 08.018.
[16]ApolinarioD,BragaRdC.O.P.MagaldiRM,BusseAL,CamporaF,BruckiS,Lee S-YD.ShortassessmentofhealthliteracyforPortuguese-speakingadults.Rev SaudePublica2012;46:702–11.PubMedPMID:22782124.
[17]Carthery-GoulartMT,AnghinahR,Areza-FegyveresR,BahiaVS,BruckiSM, DaminA,FormigoniAP,FrotaN,GuarigliaC,JacintoAF,KatoEM,LimaEP, MansurL,MoreiraD,NobregaA,PortoCS,SenahaML,SilvaMN,SmidJ, Souza-TalaricoJN,RadanovicM,NitriniR.PerformanceofaBrazilianpopulationon thetestoffunctionalhealthliteracyinadults.RevSaudePublica2009;43: 631–8.PubMedPMID:19488667.
[18]ChangLC,HsiehPL,LiuCH.PsychometricevaluationoftheChineseversionof short-form test of functionalhealth literacy in adolescents. J Clin Nurs 2012;21:2429–37.http://dx.doi.org/10.1111/j.1365-2702.2012.04147.x. [19]Baron-EpelO,BalinL,DanielyZ,EidelmanS.ValidationofaHebrewhealth
literacytest.PatientEducCouns2007;67:235–9.PubMedPMID:17386994. DOI:10.1016/j.pec.2007.02.005.
[20]HanHR,KimJ,KimMT,KimKB.Measuringhealthliteracyamongimmigrants withaphoneticprimarylanguage:acaseofKorean Americanwomen.J ImmigrMinorHealth2011;13:253–9.PubMedPMID:20585985.Pubmed CentralPMCID:PMC3010254.DOI:10.1007/s10903-010-9366-0.
[21]GuilleminF,BombardierC,BeatonD.Cross-culturaladaptationof health-relatedqualityoflifemeasures:literaturereviewandproposedguidelines.J ClinEpidemiol1993;46:1417–32.PubMedPMID:8263569.
[22]BandaranayakeRC. Settingandmaintainingstandardsinmultiplechoice examinations:AMEEGuide No.37.MedTeach2008;30:836–45.PubMed PMID:19117221.DOI:10.1080/01421590802402247.
[23]RickerKL.Settingcut-scores:acriticalreviewoftheAngoffandmodified Angoffmethods.AlbertaJEducRes2006;52:53–64.PubMedPMID:11165581.
[24]Hu LtBentler PM.Cutoff criteriafor fitindexes in covariancestructure analysis:conventionalcriteriaversusnewalternatives.StructEquModeling 1999;6:1–55.http://dx.doi.org/10.1080/10705519909540118.
[25]GraingerJ,SeguiJ.Neighborhoodfrequencyeffectsinvisualwordrecognition: acomparisonoflexicaldecisionandmaskedidentificationlatencies.Percept Psychophys1990;47:191–8.PubMedPMID:2304817.
[26]WeirJP.Quantifyingtest–retestreliability usingtheintraclasscorrelation coefficientandtheSEM.JStrengthCondRes2005;19:231–40.PubMedPMID: 15705040.
[27]JordanJE,OsborneRH,BuchbinderR.Criticalappraisalofhealthliteracy indicesrevealedvariableunderlyingconstructs,narrowcontentand psycho-metric weaknesses. J Clin Epidemiol 2011;64:366–79. PubMed PMID: 20638235.DOI:dx.doi.org/10.1016/j.jclinepi.2010.04.005.
[28]SorensenK,VandenBrouckeS,FullamJ,DoyleG,PelikanJ,SlonskaZ,BrandH. ConsortiumHealthLiteracyProjectE.Healthliteracyandpublichealth:a systematicreviewandintegrationofdefinitionsandmodels.BMC Public Health 2012;12:80. PubMed PMID: 22276600. Pubmed Central PMCID: 3292515.DOI:10.1186/1471-2458-12-80.
