The prevalence and characteristics of frailty ... - BMC Geriatrics

The frailty phenotype is defined by the presence of three from the following five clinical features: weakness, slow walking speed, ... Skiptomaincontent Advertisement SearchallBMCarticles Search DownloadPDF Researcharticle OpenAccess Published:16November2018 TheprevalenceandcharacteristicsoffrailtybyfrailtyphenotypeinruralTanzania EmmaGraceLewis1,2,6,SelinaColes3,KateHoworth2,JohnKissima4,WilliamGray2,SarahUrasa5,RichardWalker1,2&CatherineDotchin1,2  BMCGeriatrics volume 18,Article number: 283(2018) Citethisarticle 6180Accesses 19Citations 6Altmetric Metricsdetails AbstractBackgroundThefrailtyphenotypeisdefinedbythepresenceofthreefromthefollowingfiveclinicalfeatures:weakness,slowwalkingspeed,unintentionalweightloss,exhaustion,andlowphysicalactivity.Ithasbeenwidelyappliedindifferentresearchandclinicalcontexts,includingacrossmanylowandmiddle-incomecountries.However,thereisevidencethattheoperationalisationofeachcomponentofthefrailtyphenotypesignificantlyaltersitscharacteristicsandpredictivevalidity,andcareisneededwhenapplyingthephenotypeacrosssettings.Thestudy’sobjectivewastooperationalisethefrailtyphenotypeinaruralTanzanianpopulationofoldercommunity-dwellingadults.MethodsConsentingadultsaged≥60 years,andresidentinfiverandomlyselectedvillagesofHaidistrictDemographicSurveillanceSite,wereeligibletoparticipateinthiscross-sectionalstudy.Fromascreenedsampleof1207olderadults,235wererandomisedandconsentedtoanassessmentoftheirfrailtystatusbythefrailtyphenotype.Trainedresearchfieldworkers(Tanzanianmedicaldoctorsandnurses)carriedoutmeasurementsandquestionnairesatlocalvillagecentresoratparticipants’homes.ResultsTheprevalenceofthefrailtyphenotype,calculatedfromcompletedatafor196participants,was9.25%(95%CI4.39–14.12)Whenmissingdatawerecountedasmeetingfrailtycriterion(i.e.missingduetoinabilitytoperformanassessment),theprevalenceincreasedto11.22%(95%CI7.11–15.32).Frailtybyphenotypecriteriawasmorecommoninolderagegroups,andwasassociatedwithself-assessedpoorhealthanddepressionsymptoms.ConclusionsFrailtycanbesuccessfullyestimatedusingthefrailtyphenotype,howevertherearechallengesinitsoperationalisationcross-culturally.Furtherworkisneededtoexplorethepotentialclinicalapplicationofthefrailtyphenotypeinsuchsettings. PeerReviewreports BackgroundFrailtywasdescribedin2001,asbeingasyndromiccondition,distinctfrom,butoverlappingwith,disabilityandcomorbidity[1].Thesyndromicapproachtofrailtyassertsthatfrailtycanbedescribedthroughthemeasurementoffivephysicalfeatures,andthatfrailtyispresentwhenthreeofthesefeaturesarepresent,andpre-frailtywhen1–2featuresarepresent.Thefivephysicalfeaturesare:unintentionalweightloss,weakness(lowhandgrip-strength),exhaustion,slowwalkingspeedandlowphysicalactivity[1,2].Thephenotypemodelisbasedonaproposedtheoretical“cycleoffrailty”whichhypothesisesthatclinicalfeaturesmeasurableinfrailty,anditsreducedphysiologicreserve,areduetoacycleoffactorsincludingdisease,chronicundernutrition,sarcopenia,andareducedrestingmetabolicrate[1,3].SincetheoriginalFrailtyPhenotype(FP)describedbyFriedetal.aspartoftheCardiovascularHealthStudy[1],thephenotypemodelhasbeenadaptedandoperationalisedaccordingtodifferingstudydesigns,methodologiesandsettings.Inlowandmiddle-incomecountries(LMIC),arecentsystematicreviewfoundthirtysixstudieshadusedtheFPtoresearchfrailty[4].TheneedforculturaladaptionofthephenotypemodelinLMICsettings,suchasTanzania,isparticularlyimportantduetoissuessuchasresourceconstraints,lowerbackgroundeducationlevelsandsocietalhealthandfunctionalnorms.HowtheFPisoperationalisedimpactsontheprevalenceoffrailty,anditspredictiveability[5].ArecentstudyhasdemonstratedthisphenomenonwhenestimatingtheprevalenceoffrailtyamongolderadultsinruralAgincourt,SouthAfrica,whereeachofthenineiterationsoftheFPconstructed,gavecorrespondingestimatesofprevalencevaryingbetween3.0and13.2%[6].However,thebenefitsoftheFPincludeitsrelativeeaseofapplication,makingitapotentialtoolforscreening.TheFPcouldbeappliedpriortoanyclinicalassessmentservingtohighlightmajorproblemsoridentifythosemostinneed[7,8].Anotherbenefittothephenotypeapproachistheminimalamountofdatarequiredforitscalculation,particularlywhencomparedwiththedeficitaccumulationapproachtofrailty(frailtyindex),whichrequiresaminimumof20deficitstobecountedinordertoproducestableestimates[9,10].Therefore,thestudy’saimwastooperationaliseandmeasuretheprevalenceoftheFPinaruralTanzanianpopulationofoldercommunity-dwellingadults.MethodsAtotalsampleof1207adults,livinginfiverandomlyselectedvillagesinHaidistrictDemographicSurveillanceSite(DSS),andaged≥60 yearswerescreenedusingtheBriefFrailtyInstrumentforTanzania(B-FIT)tool[11].Thisshortquestionnaire,previouslydevelopedbyourresearchteam,wasusedtocategoriseparticipantsintonon-frail,pre-frailorfrail[11].Arandomselectionof79(8.9%)ofnon-frail,120(42.1%)ofpre-frailand37(94.7%)offrailformedafrailty-weightedcohortof236whounderwentComprehensiveGeriatricAssessments(CGA),followingtherecommendationsoftheBritishGeriatricsSociety[12].Onecasewasexcludedduetolargeamountsofmissingdata,leaving235olderadults,whowereassessedforfrailtyaccordingtotheFP,intheirhomesorataconvenientlocalcentre.OurconstructionoftheHaidistrictDSSFP,iscomparedwithFried’sFPinTable 1.InconstructingaFPinoursetting,weaimedtokeepascloseaspossibletoFried’sFP,sothefivemeasuredcomponentsoffrailtywereoperationalisedthus: Weakness:Handgripstrength(HGS)wasmeasuredwithparticipantssittinguprightwiththearminflexionat90degrees.Threemeasurementswereaveragedfromtheparticipant’sdominanthandusingaJAMARHydraulicHandDynamometer(ModelJ000105,LafayetteInstruments,Lafayette,IN,USA). Walkingspeed:Participantswereaskedtowalkinastraightline,adistanceof4.5 m(15 ft).Thedistancewasmeasuredoutonaflatfloorsurfaceusingarigidtapemeasure.Thetypeoffloorsurfacevaried,forexampleparticipantswhowereassessedattheirhomeswereoftenassessedwalkingoutsideduetolimitedspaceindoors.However,ifthesurfacewasslipperyduetorecentrain,oruneven,asurfaceindoorswasfound.Atthelocalassessmentcentres(churchbuildings,dispensariesandlocalschools)smoothconcretefloorswereavailable.Walkingspeedwasnotadjustedforheightorsex.Participantswalkedintheirusualfootwearandwerepermittedtouseanywalkingaids. Exhaustion:Apositiveresponsewascounted,toeitherofthetwoCentreforEpidemiologicalStudiesDepressionscale(CES-D)statements[13]:“IfeltthateverythingIdidwasaneffort”or“Icouldnotgetgoing”.ThesequestionsweretranslatedverbatimtoSwahili,andparticipantswereaskedtogradehowofteninthepastweektheyhadfeltthisway. Weightloss:Participantswereasked“Haveyoulostweightduringthelast3months?”withtheoptionofanswering“Weightlossgreaterthan3kg”,“weightlossbetween1and3kg”,“noweightloss”or“doesnotknow”.ThisvariablewasturnedfromacategoricaltoabinaryvariableforanalysisasdescribedinTable1.BMIwascalculatedfromweightmeasuredusingMicrolifeDiagnosticWS80digitalweighingscalesandheightusingtheMarsdenLeicesterHeightMeasure.Giventhatnoserialmeasurementswerepossible,self-reportedweightlosswaspreferredasameasureofunintentionalweightloss. Lowphysicalactivity:TheInternationalPhysicalActivityQuestionnaire(IPAQ),[14]wasusedtorecordparticipants’estimationsoftheirphysicalactivityoverthepreceding7 days.TheIPAQhasbeenusedwidely,andinsimilarstudiesofolderadults,forexampletheIbadanstudyofageing,fromNigeria[15].Thosewhoreportednotbeingabletocarryouttasksofmoderatephysicalactivityonanydayofthepreviousweekscoredpositivelyonthisparameter.Thiswasfelttobeappropriate,giventhatcarryingout“smallworks”inthehome(suchasusingawinnowingbasketorsharpeningknives)wasidentifiedasacommonnormforolderadultsinthiscontextwhendevelopingaculturallyadaptedinstrumentalactivitiesofdailyliving(IADLs)screeningtool[16]. Table1ComparingoperationalisationoftheHaiDSSfrailtyphenotypewithfried’sfrailtyphenotypeFullsizetableFrailtyphenotypeOtherrecordedparameters:Aquestionnairewasconductedaskingparticipantsorrespondingcloserelativesaboutsocio-demographiccharacteristicsaswellasself-reporteddiagnoses,includingHIV-infection.TheIDEAcognitive[17],andEURO-D[18],screeningtoolswereusedtoscreenforcognitivefunctionanddepressionrespectively.AshortculturallyspecificInstrumentalActivitiesofDailyLiving(IADL)toolwasused[16],wheredifficultywithanyoneofeleveninstrumentalactivities,(e.g.carryingoutsmallworksinthehome,givingadviceandpresidingoverceremonies)wasclassifiedasIADLdisability.SettingThestudywascarriedoutacrossfivevillagesinHaiDSSofruralNorthernTanzaniabetween24thFebruaryand9thofAugust2017.Thefivevillageswerestratifiedaccordingtohigh-,middle-,andlow-landlocations,duetodifferingclimate,ethnicbackgroundandagriculturalpracticesacrossthesestrata.Themajorityofthebackgroundpopulationweresubsistencefarmersreliantonnon-mechanisedfarmingpractisesandonseasonalrains.Therearetworainyseasonsperyearinthisregion,thelongrainsofFebruary–AprilandshortrainsofNovember–December.TheshortrainswhichwereexpectedinNovember2016weredisappointing,meaningthatthesecondharvestoftheyearwaspoor.Thedatacollectionperiodtookplaceduringthelongrainyseasonof2017,andintothefirstharvestseason.DataanalysisDatawerecollectedbyhand-heldtabletdevicesusingOpenDataKit(ODK)opensourcesoftware[19].CompletedformswereuploadedtoasecureODK0.2aggregateonlinedatabase.DownloadeddatawerecleanedinMicrosoftExcel2016workbooksandimportedtoStata/SE15.0foranalysis.SimpledescriptivestatisticswerecalculatedandunadjustedChi-squaredvaluesarepresentedforfrailtybyvariablesofinterest(Table 3).Thealphavaluewassetat0.05.Whencalculatingtheprevalenceoffrailty,theweightedstratificationwastakenintoaccount,usinginverseproportions.Tocalculateconfidenceintervals(CI),bootstrapping(Statacommand‘svyset’)wasusedtocontrolforclusteringbyvillageandtoadjustforthestratifiedweighting[12].Inprimaryanalysis,caseswithmissingdatanecessaryforcalculatingtheFPwereexcludedfromtheanalysis,andthus,wereassumedmissingcompletelyatrandomandnon-informative.Insecondaryanalysis,valuesformissingdatawereimputedundertheassumptionthattheywerenotmissingcompletelyatrandomandinformative.ResultsPrevalenceoffrailtyThetotalstudycohortof235peoplecomprised136females(57.9%)andtheaverageagewas74.8 years.TheprevalenceoffrailtybyHaiDSSFP,calculatedfromcompletedatafor196participants,was9.25%(95%CI4.39–14.12).ImpactofmissingdataonestimatedfrailtyprevalenceThenumberofassessmentsofwalkingspeedandgripstrengthwhichweremissingwere37and18,respectively.Interestingly,11ofthe37missingwalkingspeedassessmentswereduetoimmobilityfollowingastroke,suggestingthatsuchdatawerenotmissingcompletelyatrandom,andthatthesedatawereinformative.Overall,ofthe39participantsmissingdataforcalculationoftheirFP,36(92.3%)weredeemedfrailbyCGA[12].Insecondaryanalyses,exceptfortwoerroneouswalkingspeedrecordingswhichweredeleted,allothermissingdatawerepresumedmissingduetobeingunabletocompletetheassessment(i.e.duetofrailtyratherthanmissingatrandom),andthuscategorisedasmeetingthefrailtycriterion.TheCGA-derivedfrailtydiagnosesforparticipantswithmissingdatahavebeenprovidedassupplementaryonlinedatainAdditionalfile1:TableS1.WhentheHaiDSSFPwascalculatedusingthisapproachtheprevalenceoffrailtyincreasedto11.22%(95%CI7.11–15.32).CorrelatesoffrailtybyphenotypeThemeanwalkingspeedwas0.65 m/s(std.dev0.40,range0.04–4.29 m/s),0.72 m/sinmenand0.60 m/sinwomen.ThemeandominantHGSwas21.9Kg(Std.dev8.5,range4.3–51.3),18.92Kginwomenand26.02Kginmen.ThemeanBMIinwomenwas25.01(std.dev5.2,range13.9–47.2),andinmenwas21.21(std.dev5.5,range13.4–53.3).Table 2reportsthefrequenciesandpercentagesoffrailtycomponentsbysex.Aroundonethirdofparticipantshadnofrailtycomponents(n = 63,32.1%),whileroughlyonethirdhadonefrailtycomponent(n = 67,34.2%),andtwoparticipantsfulfilledallfivecomponentsoftheFPcriteria.Table2TheFrequenciesandpercentagesofHaiDSSfrailtyphenotypecomponentsFullsizetableFrailtybyHaiDSSFPwasassociatedwitholderagegroups,self-assessedpoorhealth,self-reportedchronicdiseasesandwithself-assessedfrailtystatus(Table3).Itisimportanttonotethatfourindividualsfrom196,disclosedtheirHIV-infectedstatus.Beingassessedathome,ratherthanatalocalcentrewasstronglyassociatedwithfrailtybyunivariateanalysis.Almosthalfofparticipantsweremarried(n = 97,49.5%),howeverfrailtywassignificantlyassociatedwithbeingwidowed,separated/divorcedorsingle.Lowlevelsofschooling,andlowliteracylevels,werealsosignificantassociationswithfrailtywith53participants,(27.0%)havingreceivednoformaleducation.Highnumbersconsideredthemselvestobeill(n = 119,60.7%),orlivingwithfrailty(n = 116,59.1%).Depressionsymptoms,asreportedbytheEURO-Dscreeningtool[18],weresignificantlyassociatedwithfrailty,aswasascoreindicativeofcognitiveimpairment,accordingtotheIDEAcognitivescreeningtool[17].Aroundafifthofparticipantswerefoundtobeunderweight,accordingtoBMI,whilealmosttwofifthswereeitheroverweightorobese(Table3),yetBMIwasnotfoundtobeassociatedwithfrailtyonunivariateanalysis.Table3DemographicandclinicalcharacteristicsofthestudysampleaccordingtoHaiDSSfrailtyphenotypestatusFullsizetableFrailty,comorbidityandADLdisabilitywerefoundtobedistinct,butoverlappingentities.ThefrequenciesandpercentagesareshowninFig. 1.Notably,almostafifthoverlappedinhavingallthree(n = 36,18.4%).Twenty(10.2%)overlappedinhavingbothdisabilityandfrailty,while12(6.1%)overlappedinhavingdualdiagnosesofcomorbidityandfrailty.Fig.1ProportionalVennDiagramIllustratingtheFrequenciesandPercentagesofComorbidity,DisabilityandFrailtybyHaiDSSFrailtyPhenotype.Legend:Frequenciesandpercentagesfromthetotalsample(n = 196)aredescribed.Comorbidityn = 59,wascategorisedasanyoneself-reporting≥2diagnosesfromthefollowing(diabetes,hypertension,stroke,cataracts,arthritis,heartdisease,respiratorydisease,HIV,TB,anaemia,depression,dementia,othermentalhealthconditions,gastro-intestinalconditions,prostatic/urinaryconditions,renalfailure,orcancer).ADLDisabilityn = 47wascategorisedasbeingunabletodoanyoftheBarthelIndex[37]activitiesofdailylivingFullsizeimageDiscussionTheprevalenceoffrailtybyfrailtyphenotypeTheprevalenceoffrailtyaccordingtotheHaiDSSFPislowerthanwasfoundbytheCGA,whichfoundafrailtyprevalenceof19.1%withinthesamestudypopulation[12].OnemajorreasonforthisisthefactthattheFPrequiresparticipantstohaveacertainleveloffunctioninginordertobeassessed.Theyshouldbemobileandabletowalk(walkingaidsarepermitted),theyshouldbeabletocomprehendandfollowcommandswellenoughtoundertakeHGStesting.ThemajorityofmissingdataforassessmentofFPwerefrompatientswhowereimmobileand/orcognitivelyimpaired,forexampleduetostrokeordementia[12].IntheoriginalFriedpaper,olderadultswithcertainclinicalcharacteristicswereexcludedbecausetheirdiagnosesmayconfoundtheresults,bymimickingfrailty[1].Forexample,thosewithahistoryofParkinson’sdisease,stroke,andpoorcognitivefunction.Thosetakingmedicationsfordepression,Parkinson’sdiseaseordementiawerealsoexcluded.Inourruralsub-SaharanAfrica(SSA)setting,excludingtheseparticipantswasnotpossible,duetolowlevelsofformaldiagnosis,limitedmedicationavailabilityandpooraccesstohealthcare.Additionally,wefeelthattheirexclusionmaysignificantlyunder-representthetrueburdenoffrailty.Thisisimportant,giventhatinEuropeandSouthAfricaithasbeenshownthatoldercommunity-dwellingadultswhocouldnotbeassessedfullyusingtheFPhadhighermortalityratesthanthosewhocouldbeassessed[6,20].Longitudinalassessmentofthiscohortwouldrevealwhetherthisfindingisthecaseinoursetting.WeaknessbyhandgripstrengthHGSvariesbyethnicityandworldregion,thuswehaveappropriatelyadjustedournormalvaluesinkeepingwiththemostup-to-datevaluesforHGSnormsinolderAfricans[21].AcomprehensivereviewofnormativeHGSvaluesfoundthatdevelopingworldregionshadsignificantlylowerHGScomparedwithdevelopedregions[22],andtakingalife-courseapproachtoageing,postulatethatfactorssuchasearlygrowthandnutritionmaybeimportantfactorsaccountingforthisdisparity.Thedifferenceingripstrength(of14.7Kginmalesand5.7Kginfemales)betweenaruralolderGhanaianstudypopulationandacomparativeDutchpopulationdisappearedwhentheauthorsstandardisedtheGhanaianstudypopulationagainsttheage-groupandsex-specificheightandBMIoftheDutchstudypopulation[23].Therefore,regionalandethnicvariationsfoundinHGSmay,tosomeextentbeaccountedforbyheightandBMI[23].ThemostimportantdeterminantsofHGSvariationgloballyisacomplexquestion.Undoubtedlygeneticfactorscontribute,howeverthereisanincreasingbodyofevidencefortheinfluenceofenvironmentalfactorssuchasearlylifesocio-economicstatusonHGSvariationinoldage[24].Thecycleoffrailtyhypothesis,statesthatchronicundernutrition(protein,energyandmicronutrientdeficiencies)isintegralinthedevelopmentandprogressionoffrailty[1,3].Indeed,dietaryproteinintakehasbeenshowntobeprotectiveagainstlossofHGS[25].Wehypothesisethatnutritionalfactors,includingalowproteinandlowmicronutrientdiet,maybeimportantinthedevelopmentoffrailtyinruralSSA,supportingfindingsfromMalawi[26].However,moredetailedassessmentsofdietaryintakeandoutcomesarerequiredinordertomakecomparisonswitholderadultsinhighincomecountry(HIC)settings.SlowwalkingspeedGiventhattherearenopublishednormsofwalkingspeedacrossSSA,weoptedtousetheslowestquintileasacut-off.Stratifyingbysexandheightdidnotsignificantlyinfluenceourfindings,butaddedtothecomplexity.TheWHOStudyofAGEingandAdultHealth(SAGE)foundarangeofwalkingspeedsbetweentheslowest,(Russiaat0.61 m/s)andthefastest,(China,at0.88 m/s)[27].Ourmeanwalkingspeedwasslow,(0.65 m/s)whencomparedwiththeSAGEcountries,andwithwalkingspeedsofHICs.Forexampleameta-analysisofninecohortstudiespredictedalifeexpectancyabovethemedianforolderadultswalkingfasterthan0.8 m/s[28].Asdiscussedinthemethods,participants’footwear,unadjustedvisualimpairments,unevengroundsurfacesandlackofappropriatewalkingaidsmayhaveledtoslowerwalkingspeeds.TheconceptofexhaustionExhaustionisasubjectiveandculturallydependentvariable.InordertoremainascloseaspossibletotheoriginalFriedFP,thesamequestionstakenfromtheCES-Ddepressionscreeningquestionnairewereused,withadirecttranslationintoSwahili.However,qualitativeexplorationofthemeaningof“exhaustion”isneeded,toensurethatthisisavalidmeasureinthissetting.ForexampletheCES-Dquestion;“Howofteninthelastweekdidyoufeelyoucouldnotgetgoing?”couldhavebeeninterpretedliterallyasnothavingaccesstoameansoftransport.Thefactthatthemajorityofourstudyparticipantsself-identifiedaslivingwithfrailty(seeTable3),isinterestinginitself,suggestingthatthetranslatedtermforfrailty;“udhaifuwawazee”orliterally“weaknessoftheolderpeople”doesn’tholdthesamenegativeconnotationsasitdoesin“westernised”cultures[29].TheAgincourtfrailtystudyfacedsimilartranslationandinterpretationchallengesinusingtheCES-Dquestions,andsubstitutedaquestiononself-reportedhealthstatus[6].UnintentionalweightlossWerecordedanyrecentself-reportedweightlossasbeingsignificant,andpresumedunintentional.InthecontextoftheHIV/AIDSepidemic,weightlossandaslimbodyhabitusmaybeassociatedwithHIV[30],whichmayhaveinfluencedourrespondents’willingnesstoadmittoweightloss,giventhatitmightimplyinfection.Culturalviewsonbodyhabitusareoftencounterto“western”ideas,wherebybeingoverweightmaybeassociatedwithhealth,wealthandsocialstatus[31].Indeed,itwouldbeveryunusualforsomeoneinthissettingtoloseweightintentionally.Giventhatsubsistencefarmingisthemainsourceoffoodandhouseholdincome,thefactthatthelastshortrainyseasonhadledtoapoorharvestmayhaveinfluencedsomeofthesurveyresponses,particularlyinrelationtofoodintakeandweightloss.VeryfewstudiesofnutritionalstatusofolderadultsinSSAhavebeenconductedrurallyandveryfewarelongitudinal[32].However,inruralGabonacross-sectionalstudyshoweda25.6%prevalenceofunderweightbyBMIamongolderadults,whichincreasedto28.2%intherainyseason[33].InoursamplethemeanBMIwas25.0forwomenand21.2formen,with20.4%(n = 40)categorisedasunderweight.Despiteaglobalincreaseinobesityitmaybethatinsuchsettings,olderpeopleareatanincreasedriskfromundernutrition,duetofoodandincomeinsecurity.Across-sectionalstudyconductedinurbanKenya,UgandaandTanzaniadescribedthemajorityofolderadultsasbeing“nutritionallyvulnerable”duetoproblemsaccessingfood[34].Asaconsequence,therelationshipbetweenfrailtyandchronicundernutritionislikelytobestrongerinruralSSAthanmightbeseeninHICsandwarrantsfurtherinvestigation.LowphysicalactivityWetookinabilitytoconductmoderatelydifficultphysicalactivityonanydayoftheweekasacriterionmarkerforfrailty.Strengthsofthismethodarethatolderadultsinthissettingarefrequentlyexpectedtotakepartinactivitiessuchassweepingandgardeningonadailybasis[16].However,themeasureisconfoundedbycognitiveimpairment,inasimilarmannertotheMinnesottaLeisureTimePhysicalActivityQuestionnaire,inthatactivitiessuchastennisrequirehighlevelsofcognitivefunctioning,asarerequiredforcommonIADLs[8].Giventhatwedidnotexcludeolderadultswithcognitiveimpairmentfromourstudysample,thereisthepossibilitythatthiscomponentmayhavebeenmeasuringtheirabilitytoperformIADLs,asassessedbytheIDEA-IADLquestionnaire[16].Takingabroadermultidimensionalapproachtofrailtyassessmenthowever,thisisacademic,ifwhatmattersistheolderadult’sfunctionalcapacity.StudylimitationsThetwoparameterswherethesedata’sobjectivitycouldbeimprovedareinthemeasurementofweightlossandcalculationoflowenergyexpenditure.AnobjectivemeasurementofweightlosswouldhavebeenbeneficialinprovidingamoreaccurateestimateofthiscomponentoftheHaiDSSFP,butwouldhaveinvolvedrepeatedhomevisitstoremoteruralareasacrossdifficultterrain,particularlyintherainyseason.OnemainweaknessofthisstudyisthatwewereunabletomeasureanestimatedenergyexpenditurebasedondatafromtheIPAQduetotechnicaldifficultiesrecordingthesedata.However,theIPAQisthoughttohavebetterreliabilityinurbanascomparedtoruralsamples,anditmayhavelimitedvalidityinruralandlowliteracysettingssuchasours[14].ThereisaneedformoreresearchintothephysicalactivityofolderadultsinruralSSA(e.g.usingaccelerometers)inordertobetterassesslowphysicalactivityinthissetting,whereitispresumedthatolderadultsremainmorephysicallyactivethanmightbeseeninHICs,whereretirementfromworkisthenorm.MoredataonthedietaryintakeofolderadultsinthissettingwouldhelpjustifytheuseoftheFPasaconceptualmodel.Manystudiesusefooddiariestomonitornutritionalstatus,butthiswouldbedifficultinthissetting,giventhelowliteracyofourstudiedpopulation.Frailtyisconsideredanimportantnon-AIDSconditioninHIV-infectedindividuals[35].OneweaknessofthisstudyisthatHIVstatuswasnottested,ratherwereliedonself-disclosure.Consequently,wemayhaveunderestimatedtheprevalenceofHIVinfectionasonlyfourparticipantsdisclosedtheirHIV-infectionstatus.HIVprevalenceinourbackgroundpopulationisunknownbutlikelytobelowbasedonanonymisedtestingconductedforastrokeriskfactorsstudyinHaidistrictbetween2003and6[36],yetthismaylimitthegeneralisabilityofourfindingstootherregionsofSSAwithahigherbackgroundprevalenceofHIVinfection.ForexampletherecentAgincourtfrailtystudy,reportedthat21%aged≥40 yearswereknowntobeHIV-infected[6].AlimitationofusingtheFPtoidentifyfrailtyinthiscontext,isitsfocusonphysicalfrailty,excludingotherdomainsofassessment,suchassocialandeconomicdomains.ParticularlyinthisruralSSAsetting,olderadults’frailtystatus,aswellastheoutcomesoffrailty,maybecloselyassociatedwiththeirsocialandeconomicresources.Furtherstudyisrequiredinordertoinvestigatetheimpactofthesedomains.StudystrengthsThisstudyaddssubstantiallytothelimitedexistingbodyofknowledgeonfrailtyinolderrural-dwellingAfricans.TheFPhasbeensuccessfullyappliedacrosslanguageandculturalbarrierstoproduceameaningfulmeasureclosetotheoriginalphenotypicmodel.Onesignificantstrengthofthisstudyisinitshypothesisgeneration.Asaresultofourfindingsofhighlevelsofself-reportedweightlossandunderweight,weproposethatchronicundernutritionduetofoodinsecurityandsocio-economicfactorsmaybeimportantinthecyclicaldeteriorationofenergyandphysiologicalreservedescribedbythe“cycleoffrailty”framework,anddeservesfurtherinvestigationinthissetting.ConclusionThestudydemonstratesthattheFPcanbesuccessfullyoperationalisedinruralTanzaniatoestimatefrailtyprevalence.TheclinicalapplicationoftheFPinruralSSAisunclear.JAMARdynamometersareexpensiveandsuchtechnicalspecialistequipmentisunlikelytobeavailableforuseroutinelyasascreeningtool.Additionally,iftheFPweretobeappliedclinically,referralforfurtherspecialistcarewouldnotbepossiblegiventheregionallackofhumanresourcesforgeriatricmedicine[37].FurtherworkassessingthenutritionalstatusofolderadultsinruralSSAwouldhelptocharacterisetheassociationbetweenchronicundernutritionandfrailtyinthissetting.Additionally,futureresearchshouldexploretheoptimalpracticalapplicationoftheFP,particularlywhetheratask-shiftingapproachtoitsdeliverycouldbeemployedinordertoreachthemostvulnerablerural-dwellingolderAfricans. AbbreviationsB-FIT: BriefFrailtyInstrumentforTanzania BMI: BodyMassIndex CES-D: CentreforEpidemiologicalStudiesDepressionScale CGA: ComprehensiveGeriatricAssessment DSS: DemographicSurveillanceSite FP: FrailtyPhenotype HGS: Hand-gripstrength HICs: High-incomecountries IADL: InstrumentalActivitiesofDailyLiving IDEA: IdentificationandInterventionforDementiainElderlyAfricans IPAQ: InternationalPhysicalActivityQuestionnaire ODK: OpenDataKit SSA: sub-SaharanAfrica WHOSAGE: WorldHealthOrganizationStudyofAGEingandadulthealth ReferencesFriedLP,TangenCM,WalstonJ,NewmanAB,HirschC,GottdienerJ,SeemanT,TracyR,KopWJ,BurkeG,etal.Frailtyinolderadults:evidenceforaphenotype.JGerontolABiolSciMedSci.2001;56(3):M146–56.CAS  Article  GoogleScholar  FriedLP,FerrucciL,DarerJ,WilliamsonJD,AndersonG.Untanglingtheconceptsofdisability,frailty,andcomorbidity:implicationsforimprovedtargetingandcare.JGerontolABiolSciMedSci.2004;59(3):255–63.Article  GoogleScholar  FriedLP,WalstonJ:Frailtyandfailuretothrivein:GeriatricMedicineandGerontology.EditedbyHazzardWR,BlassJP,EttingerWHJr,HalterJB,OuslanderJ,5edition(1July2003)ednNewYork:McGraw-HillMedical;2003:1487–1502.GrayWKRJ,McGuireJ,DewhurstF,ElderV,WeeksJ,WalkerRW,DotchinCL.Frailtyscreeninginlow-andmiddle-incomecountries:asystematicreview.JAmGeriatrSoc.2016;64(4):806–23.Article  GoogleScholar  TheouO,CannL,BlodgettJ,WallaceLM,BrothersTD,RockwoodK.Modificationstothefrailtyphenotypecriteria:systematicreviewofthecurrentliteratureandinvestigationof262frailtyphenotypesinthesurveyofhealth,ageing,andretirementinEurope.AgeingResRev.2015;21:78–94.Article  GoogleScholar  PayneCF,WadeA,KabudulaCW,DaviesJI,ChangAY,Gomez-OliveFX,KahnK,BerkmanLF,TollmanSM,SalomonJA,etal.PrevalenceandcorrelatesoffrailtyinanolderruralAfricanpopulation:findingsfromtheHAALSIcohortstudy.BMCGeriatr.2017;17(1):293.Article  GoogleScholar  CesariM,GambassiG,vanKanGA,VellasB.Thefrailtyphenotypeandthefrailtyindex:differentinstrumentsfordifferentpurposes.AgeAgeing.2014;43(1):10–2.Article  GoogleScholar  TheouO,RockwoodK.Comparisonandclinicalapplicationsofthefrailtyphenotypeandfrailtyindexapproaches.In:TheouO,RockwoodK,editors.FrailtyinAgingBiological,ClinicalandSocialImplications.,vol.41.Basel:Karger;2015.p.74–84.Chapter  GoogleScholar  RockwoodK,MitnitskiA.Frailtyinrelationtotheaccumulationofdeficits.JGerontolABiolSciMedSci.2007;62(7):722–7.Article  GoogleScholar  SearleSD,MitnitskiA,GahbauerEA,GillTM,RockwoodK.Astandardprocedureforcreatingafrailtyindex.BMCGeriatr.2008;8:24.Article  GoogleScholar  GrayWK,OregaG,KisoliA,RogathiJ,PaddickSM,LongdonAR,WalkerRW,DewhurstF,DewhurstM,ChaoteP,etal.IdentifyingfrailtyanditsoutcomesinolderpeopleinruralTanzania.ExpAgingRes.2017;43(3):257–73.Article  GoogleScholar  LewisEG,WoodG,HoworthK,ShahB,MulliganL,KissimaJ,DotchinC,GrayW,UrasaS,WalkerR.Prevalenceoffrailtyinoldercommunity-dwellingTanzaniansaccordingtocomprehensivegeriatricassessment.JAmGeriatrSoc.2018.OrmeJG,ReisJ,HerzEJ.FactorialanddiscriminantvalidityoftheCenterforEpidemiologicalStudiesDepression(CES-D)scale.JClinPsychol.1986;42(1):28–33.CAS  Article  GoogleScholar  CraigCL,MarshallAL,SjostromM,BaumanAE,BoothML,AinsworthBE,PrattM,EkelundU,YngveA,SallisJF,etal.Internationalphysicalactivityquestionnaire:12-countryreliabilityandvalidity.MedSciSportsExerc.2003;35(8):1381–95.Article  GoogleScholar  GurejeO,OladejiBD,AbionaT,ChatterjiS.ProfileanddeterminantsofsuccessfulagingintheIbadanstudyofageing.JAmGeriatrSoc.2014;62(5):836–42.Article  GoogleScholar  CollingwoodCPS-M,KisoliA,DotchinCL,GrayWK,MboweG,MkendaS,UrasaS,MushiD,ChaoteP,etal.Developmentandcommunity-basedvalidationoftheIDEAstudyinstrumentalactivitiesofdailyliving(IDEA-IADL)questionnaire.GlobHealthAction.2014;7(1):25988.Article  GoogleScholar  PaddickSM,GrayWK,OgunjimiL,LwezualaB,OlakehindeO,KisoliA,KissimaJ,MboweG,MkendaS,DotchinCL,etal.ValidationoftheidentificationandinterventionfordementiainelderlyAfricans(IDEA)cognitivescreeninNigeriaandTanzania.BMCGeriatr.2015;15:53.Article  GoogleScholar  GuerraM,FerriC,LlibreJ,PrinaAM,PrinceM.PsychometricpropertiesofEURO-D,ageriatricdepressionscale:across-culturalvalidationstudy.BMCPsychiatry.2015;15:12.Article  GoogleScholar  BrunetteW,SundtM,DellN,ChaudhriR,BreitN,BorrielloG.Opendatakit2.0:expandingandrefininginformationServicesforDevelopingRegions.In:ACMHotMobile’13;2013.http://www.hotmobile.org/2013/papers/full/2.pdf.RavindrarajahR,LeeDM,PyeSR,GielenE,BoonenS,VanderschuerenD,PendletonN,FinnJD,TajarA,O'ConnellMD,etal.Theabilityofthreedifferentmodelsoffrailtytopredictall-causemortality:resultsfromtheEuropeanmaleagingstudy(EMAS).ArchGerontolGeriatr.2013;57(3):360–8.Article  GoogleScholar  LeongDP,TeoKK,RangarajanS,KuttyVR,LanasF,HuiC,QuanyongX,ZhenzhenQ,JinhuaT,NoorhassimI,etal.Referencerangesofhandgripstrengthfrom125,462healthyadultsin21countries:aprospectiveurbanruralepidemiologic(PURE)study.JCachexiaSarcopeniaMuscle.2016;7(5):535–46.Article  GoogleScholar  DoddsRM,SyddallHE,CooperR,KuhD,CooperC,SayerAA.Globalvariationingripstrength:asystematicreviewandmeta-analysisofnormativedata.AgeAgeing.2016;45(2):209–16.Article  GoogleScholar  KoopmanJJ,vanBodegomD,vanHeemstD,WestendorpRG.Handgripstrength,ageingandmortalityinruralAfrica.AgeAgeing.2015;44(3):465–70.Article  GoogleScholar  HurstL,StaffordM,CooperR,HardyR,RichardsM,KuhD.Lifetimesocioeconomicinequalitiesinphysicalandcognitiveaging.AmJPublicHealth.2013;103(9):1641–8.Article  GoogleScholar  McLeanRR,ManganoKM,HannanMT,KielDP,SahniS.DietaryproteinintakeisprotectiveagainstlossofgripstrengthamongolderadultsintheFraminghamoffspringcohort.JGerontolABiolSciMedSci.2016;71(3):356–61.CAS  Article  GoogleScholar  ChilimaDM,IsmailSJ.NutritionandhandgripstrengthofolderadultsinruralMalawi.PublicHealthNutr.2001;4(1):11–7.CAS  Article  GoogleScholar  CapistrantBD,GlymourMM,BerkmanLF.Assessingmobilitydifficultiesforcross-nationalcomparisons:resultsfromtheWorldHealthOrganizationstudyonglobalAGEingandadulthealth.JAmGeriatrSoc.2014;62(2):329–35.Article  GoogleScholar  StudenskiS,PereraS,PatelK,RosanoC,FaulknerK,InzitariM,BrachJ,ChandlerJ,CawthonP,ConnorEB,etal.Gaitspeedandsurvivalinolderadults.JAMA.2011;305(1):50–8.CAS  Article  GoogleScholar  Britainthinks:Frailty:LanguageandPerceptionsAreportpreparedbyBritainThinksonbehalfofAgeUKandtheBritishGeriatricsSocietyIn.;2015.Matoti-MvaloT,PuoaneT.PerceptionsofbodysizeanditsassociationwithHIV/AIDS.SAfrJClinNutr.2011;24(1):40–452011.Article  GoogleScholar  RenzahoAMN.Fat,richandbeautiful:changingsocio-culturalparadigmsassociatedwithobesityrisk,nutritionalstatusandrefugeechildrenfromsub-SaharanAfrica.Health&Place.2004;10(1):105–13.Article  GoogleScholar  JésusP,GuerchetM,PilleronS,FayemendyP,MaximeMouangaA,MbelessoP,PreuxPM,DesportJC.Undernutritionandobesityamongelderlypeoplelivingintwocitiesofdevelopingcountries:prevalenceandassociatedfactorsintheEDACstudy.ClinNutriESPEN.2017;21:40–50.Article  GoogleScholar  BlaneyS,BeaudryM,LathamM,ThibaultM.NutritionalstatusanddietaryadequacyinruralcommunitiesofaprotectedareainGabon.PublicHealthNutr.2009;12(10):1946–59.CAS  Article  GoogleScholar  CheserekMJ,WaudoJN,TuitoekPJ,MsuyaJM,KikafundaJK.NutritionalvulnerabilityofolderpersonslivinginurbanareasofLakeVictoriaBasininEastAfrica:acrosssectionalsurvey.JNutrGerontolGeriatr.2012;31(1):86–96.Article  GoogleScholar  BrothersTD,KirklandS,GuaraldiG,FalutzJ,TheouO,JohnstonBL,RockwoodK.Frailtyinpeopleagingwithhumanimmunodeficiencyvirus(HIV)infection.JInfectDis.2014;210(8):1170–9.Article  GoogleScholar  WalkerRW,JusabaniA,ArisE,GrayWK,UnwinN,SwaiM,AlbertiG,MugusiF.StrokeriskfactorsinanincidentpopulationinurbanandruralTanzania:aprospective,community-based,case-controlstudy.LancetGlobHealth.2013;1(5):e282–8.Article  GoogleScholar  DotchinCL,AkinyemiRO,GrayWK,WalkerRW.Geriatricmedicine:servicesandtraininginAfrica.AgeAgeing.2013;42(1):124–8.Article  GoogleScholar  DownloadreferencesAcknowledgementsWegratefullyacknowledgetheNewcastleUniversityMresstudentswhoassistedwithdatacollection:GretaWood,LouiseWhittonandHarryCollin.WewouldliketothanktheTanzaniandatacollectionteam:AloyceKisoli,AntusaJohnKissima,PaulinaEliasTukay,Dr.JoyceMkodo,Dr.DeborahMdegella,Dr.AliMohammedAli,Dr.FrancisZerdandDr.AllyMohamedImani.ThankstoJaneRogathi,LeilaMwakipundaandLucyMarikifortheirexcellenttranslationwork.WearegratefulforthetechnicalsupportofKilimanjaroClinicalResearchInstitute,KilimanjaroChristianMedicalCentre,andHaiDistrictHospital.Wewouldliketoacknowledgetheimportantroleofthevillagechairmen,enumeratorsandten-cellleaders.Thanksofcourse,gotoalloftheolderadultsandtheirrelativeswhoparticipated. Funding NorthumbriaHealthcareNHSFoundationTrustprovidedthefirstauthor’ssalarythroughaTeachingandResearchFellowGrant.Thesponsorhadnoroleinthedesign,methods,subjectenrolment,datacollection,analysisorpreparationofthispaper. Availabilityofdataandmaterials Thedatasetsusedduringthecurrentstudyareavailablefromthecorrespondingauthoronreasonablerequest. AuthorinformationAffiliationsInstituteofHealthandSociety,FacultyofMedicalSciences,NewcastleUniversity,NewcastleuponTyne,UKEmmaGraceLewis, RichardWalker & CatherineDotchinNorthumbriaHealthcareNHSFoundationTrust,NorthTynesideGeneralHospital,NorthShields,UKEmmaGraceLewis, KateHoworth, WilliamGray, RichardWalker & CatherineDotchinTheMedicalschool,FacultyofMedicalSciences,NewcastleUniversity,NewcastleuponTyne,UKSelinaColesHaiDistrictHospital,BomaNg’ombe,Hai,Kilimanjaro,TanzaniaJohnKissimaKilimanjaroChristianMedicalCentre,Moshi,Kilimanjaro,TanzaniaSarahUrasaEducationcentre,NorthTynesideGeneralHospital,RakeLane,NorthShields,NE298NH,UKEmmaGraceLewisAuthorsEmmaGraceLewisViewauthorpublicationsYoucanalsosearchforthisauthorin PubMed GoogleScholarSelinaColesViewauthorpublicationsYoucanalsosearchforthisauthorin PubMed GoogleScholarKateHoworthViewauthorpublicationsYoucanalsosearchforthisauthorin PubMed GoogleScholarJohnKissimaViewauthorpublicationsYoucanalsosearchforthisauthorin PubMed GoogleScholarWilliamGrayViewauthorpublicationsYoucanalsosearchforthisauthorin PubMed GoogleScholarSarahUrasaViewauthorpublicationsYoucanalsosearchforthisauthorin PubMed GoogleScholarRichardWalkerViewauthorpublicationsYoucanalsosearchforthisauthorin PubMed GoogleScholarCatherineDotchinViewauthorpublicationsYoucanalsosearchforthisauthorin PubMed GoogleScholarContributionsEGL,KH,SC,WG,CD,RWandSUwereinvolvedinstudyconceptanddesign.AuthorsEGL,KH,SC,andJKassistedintheenrolmentandassessmentofparticipants.EGL,KH,WG,CDandRWwereinvolvedintheanalysisandinterpretationofdata,andallauthorsassistedinthepreparationandapprovalofthemanuscript.CorrespondingauthorCorrespondenceto EmmaGraceLewis.Ethicsdeclarations Ethicsapprovalandconsenttoparticipate ThisstudywasapprovedbytheNationalInstituteofMedicalResearch,Dar-es-Salaam,Tanzania,KilimanjaroChristianMedicalCollegeResearchEthicsandReviewCommittee,Tanzania,andNewcastleUniversityResearchEthicsCommittee,UnitedKingdom.Allparticipantswereprovidedwithwrittenandverbalinformationaboutthestudypriortoconsentinginwriting,eitherprovidingasignatureorthumbprint.Closefamilymembersprovidedtheirwrittenassentonbehalfofparticipantslackingcapacitytoconsent. Consentforpublication Notapplicableasnoindividualdataisreported,alldatahasbeenanonymisedandparticipantsarenotidentifiable. Competinginterests Theauthorsdeclarethattheyhavenocompetinginterests. Publisher’sNote SpringerNatureremainsneutralwithregardtojurisdictionalclaimsinpublishedmapsandinstitutionalaffiliations. Additionalfile Additionalfile1:TableS1.Thefrailtystatusanddiagnosesofthe39participantswithmissingHaiDSSFPdata.(DOCX16 kb)Rightsandpermissions OpenAccessThisarticleisdistributedunderthetermsoftheCreativeCommonsAttribution4.0InternationalLicense(http://creativecommons.org/licenses/by/4.0/),whichpermitsunrestricteduse,distribution,andreproductioninanymedium,providedyougiveappropriatecredittotheoriginalauthor(s)andthesource,providealinktotheCreativeCommonslicense,andindicateifchangesweremade.TheCreativeCommonsPublicDomainDedicationwaiver(http://creativecommons.org/publicdomain/zero/1.0/)appliestothedatamadeavailableinthisarticle,unlessotherwisestated. ReprintsandPermissionsAboutthisarticleCitethisarticleLewis,E.,Coles,S.,Howorth,K.etal.TheprevalenceandcharacteristicsoffrailtybyfrailtyphenotypeinruralTanzania. BMCGeriatr18,283(2018).https://doi.org/10.1186/s12877-018-0967-0DownloadcitationReceived:17July2018Accepted:26October2018Published:16November2018DOI:https://doi.org/10.1186/s12877-018-0967-0SharethisarticleAnyoneyousharethefollowinglinkwithwillbeabletoreadthiscontent:GetshareablelinkSorry,ashareablelinkisnotcurrentlyavailableforthisarticle.Copytoclipboard ProvidedbytheSpringerNatureSharedItcontent-sharinginitiative KeywordsPrevalenceFrailtyFrailtyphenotypeOlderadultsSub-SaharanAfrica DownloadPDF Advertisement BMCGeriatrics ISSN:1471-2318 Contactus Submissionenquiries:[email protected] Generalenquiries:[email protected]