Development and psychometric evaluation of a health-related quality of life instrument for

13 Pages·2016·0.62 MB·English

UCLA
UCLA Previously Published Works
Title
Development and psychometric evaluation of a health-related quality of life instrument for
individuals with adult-onset hearing loss.
Permalink
https://escholarship.org/uc/item/5f38g1hx
Journal
International journal of audiology, 55(7)
ISSN
1499-2027
Authors
Stika, Carren J
Hays, Ron D
Publication Date
2015-07-01
DOI
10.3109/14992027.2016.1166397

Peer reviewed
eScholarship.org Powered by the California Digital Library
University of California
International Journal of Audiology
ISSN: 1499-2027 (Print) 1708-8186 (Online) Journal homepage: http://www.tandfonline.com/loi/iija20
Development and psychometric evaluation of
a health-related quality of life instrument for
individuals with adult-onset hearing loss
Carren J. Stika & Ron D. Hays
To cite this article: Carren J. Stika & Ron D. Hays (2016): Development and psychometric
evaluation of a health-related quality of life instrument for individuals with adult-onset hearing
loss, International Journal of Audiology, DOI: 10.3109/14992027.2016.1166397
To link to this article: http://dx.doi.org/10.3109/14992027.2016.1166397
Published online: 22 Apr 2016.
Submit your article to this journal
View related articles
View Crossmark data
Full Terms & Conditions of access and use can be found at
http://www.tandfonline.com/action/journalInformation?journalCode=iija20
Download by: [Carren Stika] Date: 22 April 2016, At: 10:29
International Journal ofAudiology 2016: 1–11
Original Article
Development and psychometric evaluation of a health-related
quality of life instrument for individuals with adult-onset hearing
loss
Carren J. Stika1 & Ron D. Hays2
1SchoolofSpeechLanguage&HearingSciences,SanDiegoStateUniversity,SanDiego,USA;2DepartmentofMedicine,DivisionofGeneral
InternalMedicine&HealthServicesResearch,UniversityofCalifornia,LosAngeles,USA
6
1
0
2
ril
p
A Abstract
2
9 2 Objective:Self-reportsof‘hearinghandicap’areavailable,butacomprehensivemeasureofhealth-relatedqualityoflife(HRQOL)for
2 individualswithadult-onsethearingloss(AOHL)doesnotexist.OurobjectivewastodevelopandevaluateamultidimensionalHRQOL
0: instrumentforindividualswithAOHL.Design:TheImpactofHearingLossInventoryTool(IHEAR-IT)wasdevelopedusingresultsof
1
] at fwoacsuscogmropulpeste,dabliyte4ra0t9uraedureltvsie(w22,–a9d1viyseoaryrseoxlpde)rwtpitahnevlariynipnugt,daengdreceosgonfitAivOeHinLteravnidewfrso.mStduidfyfesreanmtpalere:aTshoef7th3e-itUemSAf.ieRlde-stuelstts:inMsturultmitreanitt
a scalinganalysissupportedfourmulti-itemscalesandfiveindividualitems.Internalconsistencyreliabilitiesrangedfrom0.93to0.96forthe
k
Sti scales.ConstructvaliditywassupportedbycorrelationsbetweentheIHEAR-ITscalesandscoresonthe36-itemShortFormHealthSurvey,
n version2.0(SF-36v2)mentalcompositesummary(r¼0.32–0.64)andtheHearingHandicapInventoryfortheElderly/Adults(HHIE/
e HHIA)(r(cid:2)(cid:3)0.70).Conclusions:ThefieldtestprovidesinitialsupportforthereliabilityandconstructvalidityoftheIHEAR-ITfor
r
ar evaluatingHRQOLofindividualswithAOHL.FurtherresearchisneededtoevaluatetheresponsivenesstochangeoftheIHEAR-ITscales
C andidentifyitemsforashort-form.
[
y
b
d Key Words: Hearing loss;health-related qualityoflife; functioning;well-being
e
d
a
o
nl
w Hearing loss has been identified as a primary contributor to the the health of Americans, identified improvement in health-related
o
D global burden of chronic disability in the United States (U.S. quality of life (HRQOL) as one of the initiative’s four primary
Burden of Disease Collaboration, 2013). Adverse effects on goals.
physical, emotional, cognitive, behavioral, occupational, and HRQOL is a multidimensional construct that encompasses the
social function have been reported (e.g. Chia et al, 2007; Dalton effects of disease or disability on an individual’s physical,
et al, 2003; Genther et al, 2013; Helvik et al, 2013; Mulrow et al, psychological, and social functioning and well-being (Spilker,
1990).Evenwhentheactualhearinglossaccordingtoaudiometric 1996). Although a number of self-report questionnaires have been
measures is categorized as ‘mild’, some individuals report signifi- developed to assist clinicians and researchers assess the impact of
cantassociateddisablingeffects(Monzanietal,2008;Mulrowetal, hearing loss on an individual’s communication functioning and
1990; Newman et al, 1997). Hence, audiometric measures alone hearingdisability,themostcommonlyusedmeasureistheHearing
yield insufficient information about an individual’s disability Handicap Inventory for the Elderly (HHIE; Ventry & Weinstein,
(Hallberg et al, 2008). Psychosocial and environmental factors, 1982). The HHIE was originally developed to quantify the
functionalimpairments,andperceivedconsequencesofthedisabil- psychosocialeffectsofhearinglossinolderadults,butresearchers
ity mustalsobe considered (Ferransetal, 2005). have used it and its shorter screening version, the HHIE-S, as a
TheWorldHealthOrganization(WHO)emphasizedthathealth hearing-targeted measureof HRQOL.
is ‘a state of complete physical, mental, and social well-being and TheHHIEiseasilyadministeredandprovidesinformationona
notmerelyanabsenceofdiseaseandinfirmity’(WHO,1948).More broad range of social and emotional factors impacted by hearing
recently, Healthy People 2020 (U.S. Department of Health and loss but it does not capture all important aspects of HRQOL. In
HumanServices,2010),a10-yearnationalobjectiveforimproving addition, the HHIE was designed for use with older individuals
Correspondence:CarrenJ.Stika,3821FrontStreet,SanDiego,CA92103,USA.E-mail:[email protected]
(Received6August2015;revised11March2016;accepted12March2016)
ISSN1499-2027print/ISSN1708-8186online(cid:2)2016BritishSocietyofAudiology,InternationalSocietyofAudiology,andNordicAudiologicalSociety
DOI:10.3109/14992027.2016.1166397
2 C. J.Stika&R. D.Hays
Abbreviations Item generation
An initial item pool was drafted based on themes identified from
AOHL Adult-Onset HearingLoss
contentanalysisofthefocusgrouptranscripts,literaturereview,and
FDA U.S. Foodand DrugAdministration
discussions with the expert advisory panel. The seven-member
HHIE HearingHandicap Inventory fortheElderly
expertadvisorypanelwascomprisedofresearchersinthefieldsof
HHIA HearingHandicap Inventory forAdults
psychology,audiology,otolaryngology,auralrehabilitation,survey
HRQOL Health-Related Qualityof Life
instrument development and psychometrics, vocational rehabilita-
IHEAR-IT Impactof HearingLoss Inventory Tool
tion, and hearing loss advocacy. In addition to being nationally
MCS MentalComponent Summary
renowned leaders in their respective fields, the majority of these
PCS Physical ComponentSummary
individuals had AOHL themselves. Questionnaire items were
SF-36v2 36-ItemShortFormHealthSurvey,version2.0
writtenfollowingguidelinessuggestedintheliteratureforwording
anddesigningHRQOLinstruments(Mullinetal,2000).Eachitem
experiencinghearingloss.Analternativeversion,theHHIA,targets was formulated following a question stem with a five-point
younger individuals by replacing three items from the HHIE with response scale (e.g. never, rarely, sometimes, often, always).
questionsthattheauthorsdeemedtoberelevanttopersonsyounger Some items, such as those pertaining to work or intimate/sexual
than65years of age(Newman etal,1990). relationships, included a ‘not applicable’ response option. A
The aim of the present study was to develop a comprehensive timeframe of ‘during the past four weeks’ was included to anchor
measureofHRQOLtargetedatindividualswithAOHLandprovide the respondent’s evaluation of his or her experiences (e.g. ‘During
initial evaluation of its psychometric properties. This paper the past four weeks, did you hold back from participating in
6 describes the development of the instrument, including its concep- conversations and discussions because you were unsure whether
01 tual framework, item generation and item reduction, scale forma- you heard what people had said?’). If the respondent used hearing
ril 2 tion,and reliability and validity. abiadssedorohnadthaeicrodcahilleyareximpeprliaenntc,etshewyitwhetrheeinhsetarruicntgeddteoviacneswone.ritems
p
A
2 Methods
2
9 Item reduction
2 Field test instrument development
0: The drafted items were organized and listed under headings that
1 Thehearing-targetedHRQOLfieldtestinstrumentwasdesignedas
at aself-reportmeasureforindividuals21yearsandolderwithAOHL. noted the content of the intended target domain and associated
] subdomain.UsingamethoddescribedbyCunninghametal(1999)
a The development of this instrument was consistent with recom-
Stik mendedguidelinesonpatient-reportedoutcomesmeasuredevelop- fmoermrabteirnsgofretlheevaenxcpyertoafdHviRsoQryOLpaniteelmwserteoaasksepdetcoifricevpieowputlhaetiolinst,
n ment issued by the U.S. Food and Drug Administration (FDA,
e and(1)judgeeachitemintermsofitsrelevancetotheHRQOLof
r 2009) and procedures endorsed by the International Society for
r
a an individual with AOHL, and (2) note items that seemed
C QualityofLifeResearch(Reeveetal,2013).Ethicsapprovalforthe
[ redundant, ambiguous, or lengthy, or contained jargon terms.
y study was obtained from the San Diego State University
b Feedbackfromtheexpertadvisorypanelresultedintheelimination
d Institutional Review Board (IRB) prior to any subject’s participa-
de tion inany phaseof thisresearch (IRB projectnumber:01642). ofseveralitems,aconsolidationofothers,andarewordingofafew.
a Thisprocess yielded an initial itempoolof87items.
o
nl Next, face-to-face cognitive interviews and respondent debrief-
w
o Conceptual framework ings were conducted with five adults with AOHL. For the
D
The initial seeds for the conceptual framework came from a study interviews, respondents were asked to read each item silently and
investigating the impact of hearing loss on a variety of aspects of then ‘think aloud’ as they determined their responses. Follow-up
daily life, including family relationships, social interactions and debriefing allowed for probing of respondents’ opinions regarding
activities, employment, use of assistive technology, and general the questionnaire, including its visual format, clarity, comprehen-
psychosocial well-being (Stika, 1997a,b; Trybus et al, 2004). siveness, and ease in completing. As a result of these pretesting
Thirteentwo-hourfocusgroupswereheldinsixdifferentlocations techniques,14itemswereomittedandseveralitemswerereworded
in the USA (Washington, DC; Allentown, PA; San Diego, CA; to enhance their clarity. A final set of 73 items were selected
ThousandOaks,CA;LongBeach,CA;Whittier,CA.),withatotal for inclusion in the field test. Also included were two global
of 107 individuals with hearing loss participating. To explore the questions that asked: ‘How would you rate your overall quality of
impact of hearing loss on daily life from the perspective of the life?’ and ‘To what extent does your hearing loss impact your
family member, five focus groups involving 37 hearing family qualityof life?’
members were alsoconducted. The final field test instrument, the Impact of Hearing Loss
A second component of the instrument development was a InventoryTool(IHEAR-IT),consistedoffourhypothesizedglobal
comprehensive literature search using the following keywords: domains(physical,psychological,social/interpersonal,andactivity/
hearingdisability,qualityoflife,health-relatedqualityoflife,adult participation) and 16associated subdomains.(SeeFigure 1)
hearingloss,andwell-being.Theaimoftheliteraturesearchwasto
identify existing HRQOL instruments used with individuals with
hearing loss, and to examine information obtained by these Existing measures and other variables
measures. Instruments designed for use with other disabilities and Existing(‘legacy’)measureswereincludedtoevaluatetheconstruct
health conditions were also reviewed to determine whether they validity of the new measure, including the 36-Item Short Form
mightcontainrelevantcontentareasanditemsthatcouldbeadapted Health Survey, version 2.0 (SF-36v2; Ware et al, 2000) and either
andincludedinthefieldtestinstrumentforindividualswithAOHL. theHHIEortheHHIA,dependingontheparticipant’sage.TheSF-
Development and psychometric evaluation of a HRQOL instrument 3
HRQOL for Individuals with Adult-Onset Hearing Loss
Physical Psychological Social/Interpersonal Ac(cid:2)vity/Par(cid:2)cipation
Sound & Speech Energy/Fa(cid:2)gue Family & Friends Recrea(cid:2)on/Community
Percep(cid:2)on (4 items) (7 items) Ac(cid:2)vity
(5 items) (5 items)
Speech Regula(cid:2)on Discouragement/ Social Engagement/ Access to Information/
(2 items) Depression Avoidance Communica(cid:2)on
(4 items) (8 items) (8 items)
Self-Esteem/ In(cid:2)macy Independence
Self-Competence (4 items) Conduc(cid:2)ng Ac(cid:2)vities
6 (5 items) of Daily Living
1
0 (3 items)
2
ril Frustra(cid:2)on/ Enjoyment of Music
p
A Irrita(cid:2)on (1 item)
2 (4 items)
2
9
0:2 Stress/Nervousness/ Work
1 Anxiety (7 items)
at (4 items)
]
a
k
Sti Safety/Vulnerability
n (2 items)
e
r
r
a
C
[
y Figure1. Diagram of thehypothesizeddomains and subdomainsof thefieldtesting instrument.
b
d
e 36v2isagenericmeasureofHRQOLconsistingof36questionsand hearing problem cause you difficulty hearing/understanding co-
d
a eightderivedscalesthatassessphysicalfunctioning,rolelimitations workers’?; ‘Does a hearing problem cause you to feel frustrated
o
nl due to physical health problems, role limitations due to emotional whentalkingtoco-workers,clients,orcustomer?’)(Newmanetal,
w
o problems, social functioning, emotional well-being, pain, energy/ 1990). The HHIA is scored and interpreted in a similar fashion as
D
fatigue,andgeneralhealthperceptions.Twosummaryscoresnormed theHHIE.
ontheU.S.generalpopulation(mean¼50,SD¼10)arecalculated:a Questions were included in the field test to gather information
physicalcomponentsummary(PCS)scoreandamentalcomponent about the respondent’s age, education, income, additional medical
summary (MCS) score. The SF-36v2 was selected because of its conditions, hearing loss, use of hearing aids and/or cochlear
widespread use in medical and healthcare research, including implants, and satisfaction with benefits obtained from assistive
researchinvolvingindividualswithdisabilities. hearing technologies if used. Also included were items that
The HHIE is a 25-item self-assessment measuring perceived addressed personal and environmental factors that might influence
psychosocial consequences of hearing loss for adults age 65 years thefunctionalperformanceoftheindividualwithhearingloss(e.g.
andolder.Therespondentisaskedtoanswer‘yes,’‘sometimes,’or ‘My family makes special effort to ensure that I am following
‘no’ to each question, with answers scored as 4, 2, or 0, conversations and discussions’; ‘I am uncomfortable asking others
respectively. Two scale scores are calculated: (1) emotional (13 to make special effort for me because of my hearing loss.’). In
items,0–52possiblerange),and(2)social/situational(12items,0– addition,potentiallypositiveoutcomesofexperiencinghearingloss
48possible range). A 25-item overallHHIE score (0–100 possible during adulthood were addressed (e.g. ‘As a result of my hearing
range)isalsocalculatedandclassifiedintothreecategoriesofself- loss,Ihaveagreaterappreciationforthechallengesfacedbypeople
perceived hearing handicap: a total score of (cid:4)16 indicates no whohavedisabilities.’).Hence,throughoutconstructionofthefield
handicap; 17–42 suggests mild to moderate handicap, and442 is test instrument, careful attention was given to broadening the
consideredasignificanthandicap.TheHHIAwasdesignedforuse conceptualizationof‘hearingdisability’beyondatraditionaldeficit
with adults younger than 65 years of age and is identical to the model and incorporating elements consistent with the WHO
original HHIE except that it includes the substitution of three International Classification of Functioning, Disability, and Health
questions which focus on occupational and social situations more (ICF) theoretical framework (2001), whereby environmental and
commonly experienced by younger adults (i.e. ‘Does a hearing personal factors that influence the disabling process are also
problem cause you difficulty in the movies or theater?’; ‘Does a considered (Haysetal, 2002;Tate & Pledger,2003).
4 C. J.Stika&R. D.Hays
Sample comprisethescale.Thisyieldedscaleswithapossiblerangeof0to
Individuals invited to participate were 21 years or older, with an 100,with higher scores indicatingbetter HRQOL.
onset of diagnosed hearing loss at 18 years of age or older. Theworkscaleitemswerenotincludedinthemultitraitscaling
Excludedwereindividualswhosehearinglosswasdiagnosedprior analysisduetotherelativelysmallsubsetofparticipantsindicating
to the age of 18, persons with serious health conditions or other thattheywerecurrentlyemployed(n¼162;40%).Similarly,items
sensory impairments that could potentially affect their daily from the intimacy/sexual relations scale were excluded from this
functioning, and individuals unable to read and understand analysis, as about one-third of thesample indicated that they were
English.Thereasonforexcludingindividualswithaknownhearing not in an intimate or sexual relationship. The psychometric
lossonsetpriortoadulthoodisthathearinglossdiagnosedlaterin properties for these two scales are discussed separately from the
lifetypicallyresultsinaverydifferentsetofissuesandadjustment multitrait scalinganalysis.
demands than when it occurs during childhood (Thomas, 1984). Contentvalidityreferstotheextenttowhichanitemreflectsthe
Individuals who are diagnosed with hearing loss during adulthood breadth and depth of the construct of interest (Hays & Revicki,
havealreadyestablishedasolidfoundationincommunicationskills, 2005). Content validity was enhanced by having members of the
a sense of self and social identity, vocational placement, and expert advisory panel rate the extent to which items, scales, and
financialresponsibilities.Moreover,theyhavecometorelyontheir subscalesreflectedcommonexperiencesassociatedwithAOHLand
abilitytohearinordertofunctionoptimallyintheirpersonallives HRQOL. In addition, extensive use of focus groups and cognitive
and socialrelationships, andatwork. interview methodology during the development of the instrument
Subjects were recruited both locally and nationally using a maximizedits contentvalidity.
variety of procedures, including flyers distributed through private Construct validity refers to the extent to which an instrument
and hospital audiology clinics, and notices placed on internet actually measures the target construct in ways that it purports it
6
1 bulletin boards, listservs, and consumer organizations’ websites does.Constructvaliditywasevaluatedbyexaminingthestrengthof
0
2 (e.g.Hearing Loss Associationof America). thecorrelationsbetweenscoresontheIHEAR-ITscalesandscores
ril on the HHIE/HHIA, and the SF-36v2 PCS and MCS. We
p
A hypothesizedthatthenewscaleswouldshowanegativecorrelation
22 Data collection withtheHHIE/HHIAtotalscoreandapositivecorrelationwiththe
9 Questionnaire packets were mailed to individuals who indicated a MCSfromtheSF-36v2,butnotcorrelatewiththePCSfromtheSF-
2
0: desire to participate in the study, met the inclusion criterion, and 36v2. In addition, based on previous research (Chia et al, 2007;
1
at hadsignedandreturnedaninformedconsentdocument.Participants Dalton et al, 2003), we anticipated that severity of hearing loss
] were asked to provide a copy of their most recent audiogram wouldbeassociatedwithpoorerHRQOL.Further,wehypothesized
a
k conducted within the last three years to examine associations of that younger individuals and women would report worse HRQOL.
Sti audiometric test results with self-reported levels of hearing loss. Finally, we hypothesized that the IHEAR-IT scales would explain
n
e Subjects were mailed a US$20 gift card to a store of their choice uniquevariancebeyondtheHHIE/HHIEinglobalratingsofoverall
r
ar (Home Depot, Target, or Starbucks) after the research packet was quality of life and the impact of hearing loss on quality of life.
C
[ returned and judged complete, regardless of whether a copy of a Following the criteria suggested by Cohen (1992), correlation
y
b recentaudiogram wasincluded. coefficientvaluesof0.1–0.23wereregardedasasmalleffectsize,
d
e 0.24–0.36 as a medium effect size, and 0.37 or higher as a large
d
a effect size.
o
nl Psychometric analyses MultitraitscalinganalyseswereperformedusingSAS(cid:3)(Hays&
w
o Items were grouped according to hypothesized domains and Wang, 1992) while all other analyses were conducted using IBM
D subdomains based upon an a priori theoretical framework and SPSSversion 22(IBMCorp.,released 2013).
item content. The responses categories were recoded and trans-
formedlinearlytoapossiblerangeof0 to100,withhigher scores
Results
indicativeof abetter HRQOL.
Next, multitrait scaling analysis was used to evaluate item Sample characteristics
discrimination across the hypothesized scales (Hays & Fayers, Atotalof443individualsexpressedinterestinparticipatinginthe
2005). Item discrimination refers to the extent to which an item studyandsignedandreturnedtheinformedconsent;412individuals
correlates significantly higher with the scale that it is intended to completed and returned the questionnaire packet (response rate
measure than it does with any of the other scales. Item discrim- ¼93%). Three individuals who returned their completed question-
ination was supported when the correlation of an item with its naireswereineligibleforparticipationbecausetheyindicatedtheir
hypothesized scale was at least two standard errors higher than hearinglosswasdiagnosedbefore18years ofage.Approximately
correlations with other scales. If an item showed a higher halfofthestudygroup(51%)providedacopyoftheiraudiogram.
correlation with a scale other than its hypothesized scale, then the The average age of the 409 study participants was 63 years
itemwasmovedtoanotherscaleandthemultitraitscalinganalysis (range:22–91years,SD¼14.0),with66%beingwomen.Themale
wasrunagain.Itemsthatcorrelatedequallywithmultiplescalesand participants tended to be older (M¼68 years, SD¼12.8) than the
items that failed to have an item-scale correlation with their female participants (M¼61 years, SD¼14.0), t(407)¼5.32,
hypothesized scale greater than 0.35 were considered individual p50.001.Theaveragereportedageofclinicaldiagnosisofhearing
items rather than associated with a particular scale. A series of loss was 43 years for women and 52 years for men. Duration of
iterations of multitrait scaling analyses were conducted until a hearinglossfortheparticipants,basedonageattimeofdiagnosis,
satisfactory‘fit’foreachitemwasachieved.ScoresontheIHEAR- ranged from less than one year to 65 years (M¼17.3, SD¼12.7),
IT were computed by adding together item scores for each scale, with26%oftheparticipantsreportinghavinghadahearinglossfor
and then dividing the total score by the number of items that sixyearsorless.Therewasnosignificantdifferencebetweenmen
Development and psychometric evaluation of a HRQOL instrument 5
andwomenonself-reporteddurationofhearingloss,t(405)¼1.36, audiometric classifications indicated ‘fair’ agreement (Landis &
p¼.175. However, women reported obtaining a diagnosis sooner Koch, 1977), k¼0.350 (95% CI, 0.26 to 0.44), p50.0005, with
afterfirstnoticingproblemswithhearing(M¼3.6years,SD¼6.1) self-reported ratings suggesting generally greater hearing impair-
thandidmen (M¼5.4years, SD¼7.5), t(401)¼2.62,p50.01. ment than audiometric findings, z¼(cid:3)7.58, p50.0005. The study
Thesampleincludedpersonsin44states,withamajorityofthe sample was negatively skewed as assessed by Shapiro-Wilk’s test
participants (31%) from California, followed by 9% living in (p50.05),withasignificantlylargerproportionoftheparticipants
Florida.Fifty-threepercentjudgedtheirhealthtobeeither‘excellent’ describing their hearing loss as being in the severe or profound
or ‘very good.’ Additional sociodemographic information for the range.Inaddition,thewomentendedtodescribetheirhearingloss
sample, including race/ethnicity, marital status, educational level, asbeing more severethan themen(z¼(cid:3)3.26,p50.01).
employmentstatus,andincomeisprovidedinTable1. Themajorityoftheparticipantsreportedusingsomeformofan
Twenty-fourpercentdescribedtheirunaidedhearinglossintheir assistivehearingdeviceonaregularbasis,with70%reportingthat
betterearas‘profound,’33%as‘severe,’28%as‘moderate,’10% theyused either one ortwo hearing aids and 17%noting thatthey
as ‘mild,’ and 4% as ‘normal’ (i.e. unilateral). There was no usedeitheroneortwocochlearimplants.Thirteenpercentreported
significant difference between self-reported levels of hearing loss that they did not use either a hearing aid or a cochlear implant.
forthosewhoprovidedaudiograms(n¼209)andthosewhodidnot Amongthislattersubsetofindividuals,50%describedthelevelof
(n¼200), z¼(cid:3)1.39, p¼0.16. Cohen’s k for agreement between hearing loss in their better ear as ‘normal’ or ‘mild.’ Fifty-three
self-reported levels of unaided hearing loss in the better ear and percentoftheparticipantswhoindicatedusinganassistivehearing
device reported being either ‘satisfied’ or ‘very satisfied’ with
benefits derived from their hearing device, while 25% reported
Table 1. Demographic characteristics of study participants
being either ‘dissatisfied’ or ‘very dissatisfied’ (see Table 2).
6 (N¼409).
1 Reported level of satisfaction was not associated to a statistically
0
2 Characteristic n % significantdegreewithseverityofhearingloss,basedoneitherself-
ril reportsoraudiometricmeasures(r¼0.03and(cid:3)0.10,respectively).
p Gender
A
2 Male 141 34 Descriptive statistics and reliability
2 Female 268 66
29 Age,years We performed 10 iterations of multitrait scaling on questionnaires
0: 420and(cid:4)35 6 2 completed with no missing data or ‘not applicable’ responses
1
at 435and(cid:4)50 73 18 (n¼376), resulting in four subscales which we labeled: psycho-
] 450and(cid:4)65 159 39 logical/emotional (20items),social/interpersonal (14items),activ-
a
k 465and(cid:4)80 122 30 ity/community participation (8 items), and access to information/
Sti 480and(cid:4)91 49 11 communication (13 items). Item-scale correlations for five items
n Race/ethnicity
e
rr White(notofHispanicorigin) 381 93
a Table2. Audiologic characteristics ofthestudy participants.
C White(HispanicorLatino) 13 3
[
y BlackorAfricanAmerican 6 2 Characteristic n %
ed b ANsaitaivne/PAacmifeicricIaslnanodreArlaskanNative 36 11 Severityofhearinglossinthebetterear(self-report)(n¼409)
ad Maritalstatus Normal(unilateralhearingloss) 18 4
nlo Single-nevermarried 37 9 Mild 41 10
w Moderate 116 28
Marriedorlivingasmarried 238 58
Do Separated 5 1 Severe 133 33
Profound 97 24
Divorced 68 17
Idon’tknow 4 1
Widowed 60 15
Severityofhearinglossinthebetterear(audiogram)(n¼199)
Education
Normal(unilateralhearingloss)(PTA 520dBHL) 7 3
5Highschool 6 2 4
Mild(PTA ¼20–40dBHL) 37 19
Highschoolgraduate 32 8 4
Moderate(PTA ¼41–70dBHL) 93 47
Vocationalschoolorsomecollege 92 22 4
Severe(PTA ¼71–90dBHL) 39 20
Four-yearcollegedegree 123 30 4
Profound(PTA 490dBHL) 23 11
Professionalorgraduatedegree 156 38 4
Typeofhearingdeviceused(n¼409)
Employmentstatus
Nohearingaidorcochlearimplant 54 13
Workingfull-time 126 31
Onehearingaid 52 13
Workingpart-time 36 9
Twohearingaids 233 57
Unemployed 23 5
Onecochlearimplant 37 9
Retired 199 49
Twocochlearimplants 11 3
Full-timehomemaker 21 5
Onecochlearimplant,onehearingaid 22 5
Volunteerworker 4 1
Satisfactionwithassistivehearingdevicebenefits(n¼352)
Totalhouseholdincome
Verysatisfied 45 13
5US$15,000 14 3
Satisfied 142 40
US$15,000–$25,000 31 8
Neutral 76 22
US$25,001–$50,000 104 25
Dissatisfied 73 21
US$50,001–$75,000 80 20
Verydissatisfied 16 4
US$75,001–$100,000 40 10
4US$100,000 79 19 PTA ¼Pure-tone average of hearing thresholds at 500, 1000, 2000, and
4
Iprefernottosay 61 15 4000Hz.
6 C. J.Stika&R. D.Hays
showed that they did not belong in any one scale, so these were SF-36v2 component scores are also shown. As hypothesized,
treatedasstand-aloneitems.AlistoftheIHEAR-ITscalesandtheir significant positive (‘medium’ to ‘large’) correlations were found
corresponding items are presented in the supplemental materials betweenscoresonthenewscalesandscoresontheSF-36v2MCS
(Supplementary Appendix A to be found online at http:// (r¼0.32–0.64). None of the IHEAR-IT scales correlated signifi-
informahealthcare.com http://informahealthcare.com/loi/ija). cantly with the SF-36v2 PCS. Further, consistent with our
Table 3 provides mean scores, standard deviations, floor and hypotheses, significant negative (‘large’) correlations were found
ceilingeffects(scoresof0or100,respectively)forthefourmulti- between the HHIE/HHIA total score and scores on the new scales
itemscalesandfivesingleitems,andinternalconsistencyreliability (r (cid:2)(cid:3)0.70), such that higher levels of self-reported hearing
estimates for the four scales. Mean scale scores ranged from 45.1 handicap, as measured by the HHIE/HHIA, were associated with
(access to information/communication) to 69.8 (psychological/ poorer HRQOL on the IHEAR-IT. The IHEAR-IT psychological/
emotional). Floor and ceiling effects were generally small for the emotional scale correlated strongly with the SF-36v2 MCS
foursubscales,withlessthan1%ofthesampleobtainingthelowest (r¼0.64). Of lesser magnitude was the association between the
possiblescore(0)onanyofthescales,andatmost7%ofthesample access to information and communication scale and the MCS
achieving the highest possible score (100) on one scale (activity/
community participation). The five individual items had greater
floor effects, ranging from 13% (effort involved compensating for Table4. MeansandstandarddeviationsforscoresontheSF-36v2
hearing loss) to 22% (enjoyment of music), and ceiling effects, and HHIE/HHIA (n¼376).
ranging from 4% (satisfaction with ability to understand informa-
%scoringat %scoringat
tion presented inlectures and meetings) to20% (effort requiredto Scales Meanscore SD thefloor theceiling
compensate forhearing loss).
6 SF-36v2a
1 Cronbach’salphaforthemulti-itemsscalesrangedfrom0.93to
0 Physicalfunctioning 48.8 9.1 0.3 24.5
2 0.96.Threeofthefourscaleswerenegativelyskewed;theexception
pril being access to information/communication, which was positively RBooldei:lyphpyasinical 4590..15 89..93 00..53 3262..41
A skewed(0.22).Fouroftheindividualitemswerepositivelyskewed Generalhealth 50.0 8.9 0.0 4.0
22 (range 0.02to 0.39). Vitality 51.1 9.8 0.8 1.3
9 Socialfunctioning 49.2 9.3 0.0 46.0
2
0: Work and intimacy/sexual relations scales Role:emotional 48.8 9.8 0.5 52.9
] at 1 Bscoatlhe t(h4eitwemorsk) sdceamleon(s7traitteemdsh)igahndinttheernianlticmoancsyis/tseenxcuyalrerleilaabtiiolintys MPCeSntalhealth 5409..16 99..80 00..00 40..30
ka MCS 49.9 10.6 0.0 0.0
Sti (a¼0.89and 0.85,respectively), were slightly negatively skewed, HHIE/HHIA
n had a wide spread in scores, and showed small floor and ceiling Totalscoreb 53.0 23.9 0.3 1.1
re effects (see Table 3). Emotionalc 25.6 13.6 2.1 2.9
r
Ca Sociald 27.4 11.4 0.5 2.1
y [ Construct validity Descriptivestatisticsreflectscoresforparticipantsincludedinthemultitrait
b
d Table 4 shows descriptive statistics for scores obtained on the SF- scalinganalysesandnottheentirestudysample.SD¼standarddeviation.
de 36v2 and the HHIE/HHIA. Mean SF-36v2 scores were about the aTheSF-36v2scalesarenormedtonationaldataandhaveameanof50anda
oa same as fortheU.S. general population, rangingfrom 49to 51on standarddeviationof10,withlowerscoresreflectingpoorerqualityoflife
nl theT-score metric. andhigherscoresindicatingbetterfunctioning.bTheHHIE/HHIAtotalscore
w hasapossiblerangeof0to100,withhigherscoresreflectinggreaterhearing
Do Table 5 presents correlations between the IHEAR-IT four handicapandlowerscoresindicatingbetterfunctioning.cThissubscalehasa
subscales, the SF-36v2, and the HHIE/HHIA. For comparison possiblerangeof0to52;higherscoresindicategreaterhearinghandicap.
purposes,correlationsbetweentheHHIE/HHIAtotalscoreandthe dThissubscalehasapossiblerangeof0to48;higherscoresindicategreater
hearinghandicap.
Table3. Descriptive statisticsand reliability estimates fortheIHEAR-IT scales and items inthefieldtest (n¼376).
Scales/individualitems Numberofitems Meanscore SD %scoringatthefloor %scoringattheceiling Cronbach’s(cid:2)
Psychological/emotional 20 69.8 18.7 0.0 0.0 0.96
Social/interpersonal 14 56.5 22.6 0.0 0.5 0.95
Activity/communityparticipation 8 62.4 26.4 0.8 7.2 0.93
Accesstoinformationandcommunication 13 45.1 23.5 0.3 0.0 0.94
Effortcompensatingforhearingloss(Q3m) 1 53.8 32.8 12.8 19.7 NA
Stresswhentalkingonthephone(Q3p) 1 49.7 34.4 18.6 18.6 NA
Satisfactionlisteningtolectures(Q4g) 1 39.6 27.5 16.0 4.0 NA
Satisfactionattendingmovies/plays(Q4h) 1 39.5 29.6 19.1 6.1 NA
Satisfactionlisteningtomusic(Q41) 1 45.0 33.5 22.1 11.2 NA
Worka(n¼162) 7 65.1 24.2 0.0 5.0 0.89
Intimacyb(n¼239) 4 60.9 26.9 3.3 8.8 0.85
Allscalesarescoredwithapossiblerangeof0to100,withlowerscoresreflectingpoorerqualityoflifeandhigherscoresindicatingbetterfunctioning.
NA¼notapplicable;SD¼standarddeviation.
aOnlyparticipantswhoindicatedbeingactivelyemployedcompleteditemsthatcomprisetheworkscale.Responsesontheseitemswerenotincludedinthe
largermultitraitscalinganalysis.bOnlyparticipantswhoindicatedbeingcurrentlyinanintimaterelationshipcompleteditemsthatcomprisetheintimacy/
sexualrelationsscale.Responsesontheseitemswereanalysedseparatefromthelargermultitraitscalinganalysis.
Development and psychometric evaluation of a HRQOL instrument 7
(r¼0.32). Not surprisingly, the correlations of greatest strength (r¼0.12 to 0.30), with older individuals generally reporting better
emerged between the HHIE/HHIA total score and the social/ HRQOL than younger individuals. Further, both self-reports of
interpersonalandthepsychological/emotionalscalesoftheIHEAR- hearing loss level and audiometric findings showed significant
IT (r¼(cid:3)0.84 and(cid:3)0.80,respectively). relationships with the IHEAR-IT scales, including negative correl-
ations of medium strength with the psychological/emotional scale
ExaminingassociationsbetweenIHEAR-ITscalescoresand (r¼(cid:3)0.27 and (cid:3)0.29, respectively), and negative correlations of
demographic variables large magnitude with the other three subscales (r¼(cid:3)0.39
Table 6 shows correlations between select demographic variables to(cid:3)0.52).
andscoresontheIHEAR-IT.Consistentwithourhypotheses,three After controlling for age, duration of hearing loss showed
subscales were significantly associated with age (small to medium significant negative correlations of small to medium strength with
strength)(psychological/emotional,r¼0.31,p50.001;social/inter- all four IHEAR-IT scales and individual items, indicating that
personal, r¼0.20, p50.001; and access to information/communi- individualswhohadlivedwithahearinglossforlongerperiodsof
cation,r¼0.12,p50.05),butthecorrelationwasnotsignificantfor time reported higher levels of HRQOL than individuals whose
the activity/community participation scale (r¼0.02, p40.05). hearinglosshadoccurredmorerecently.Annualhouseholdincome
Similarly, age showed significant positive (small to medium and educational level correlated significantly with the activity/
strength) correlations with four of the five individual items communityparticipationscale(r¼0.16and0.11,respectively),but
notwith anyother IHEAR-IT scale orindividual item.
FemalesreportedsignificantlypoorerHRQOLthanmenontwo
Table 5. Pearson correlations of the IHEAR-IT scales with the ofthefourIHEAR-ITscales(psychological/emotionalandaccessto
SF-36v2summaryscoresandtheHHIE/HHIAtotalscore(n¼376). information),aswellasontwoofthefiveindividualitems(‘effort
6
1 requiredtocompensateforhearingloss’and‘stresswhentalkingon
0 SF-36v2 SF-36v2 HHIE/HHIA
ril 2 PCS MCS totalscore tahsesopchiaotnioentsowanasunsfmamalill,iawriptherstohne’)e.xHceopwtieovnero,fthtehestrpesnygcthhoolofgthiceasle/
p IHEAR-ITscales
A emotional scale, which was of medium strength (t(374)¼5.08,
22 SPosycciahlo/ilnotgeircpaelr/esomnoatlional (cid:3)00..0012 00..4684**** (cid:3)(cid:3)00..8840**** p50.001,d¼0.56).
29 Activity/communityparticipation 0.10 0.34** (cid:3)0.70** AseriesofmultivariateANOVAswereconductedtodetermine
0: Accesstoinformationandcommunication 0.04 0.32** (cid:3)0.75** the effect of hearing device (i.e. hearing aid versus cochlear
1
at Worka(n¼162) 0.08 0.36** (cid:3)0.65** implant)onIHEAR-ITscores.Aftercontrollingforageandgender,
] Intimacyb(n¼239) (cid:3)0.01 0.39** (cid:3)0.62** there was a statistically significant difference for hearing device
a
en Stik SSHFFH-I¼3E63/vH62vH2PICAMSCtoStalscore (cid:3)010...000409 (cid:3)(cid:3)010...004907** (cid:3)010...004407** aLsicagrmnoisbfsiact¼ahne.t8d6de0ipf)fe.enrHdeeonnwcteesvvaebrrei,atwbfoleelelsonw(Fc-(uo9pc,h3ule1na4irv)a¼irmia5pt.el6a7nF,tptae5nsdt0s.h0seh0ao1rw;inWegd-ialnkidos
rr HHIE/HHIAsocial/situationalscore 0.03 (cid:3)0.38** 0.95**
a users for scores on the IHEAR-IT subscales. Only one of the five
C HHIE/HHIAemotionalscore 0.06 (cid:3)0.49** 0.96**
[ individual items yielded a significant group difference, with
y
b PCS¼physical componentscore; MCS¼mental component score; HHIE/ individuals with cochlear implants reporting that they experienced
ed HHIA¼hearing handicap inventory for the elderly / hearing handicap significantlylesssatisfactionlisteningtomusic(M¼28.0,SD¼30)
d inventoryforadults.Correlationswithoutasuperscriptarenotsignificant
a than individuals with hearing aids (M¼46.9, SD¼33.2). Similar
o (p40.05).
nl **p50.01. ANOVAs were conducted to examine the effect of hearing device
w
o aOnlyparticipantswhoindicatedbeingactivelyemployedwereincludedin ontheHHIE/HHIAtotalscoreandsubscalescores,andalsotheSF-
D theworkscale. 36v2PCSandMCS.Findingsindicatednon-significantassociations
bOnlyparticipantswhoindicatedbeingcurrentlyinanintimaterelationship
betweenhearing device and scores oneither measure.
wereincludedintheintimacy/sexualrelationsscale.
Table6. Associationsof IHEAR-IT scalesand individual items withdemographicvariables and hearing loss.
Annual Highest Self-reportedhearing Audiometrichearing
Scales/individualitems Age income educationallevel losslevela losslevelb
Psychological/emotional 0.31** 0.08 0.04 (cid:3)0.27** (cid:3)0.29**
Social/interpersonal 0.20** 0.10 0.08 (cid:3)0.39** (cid:3)0.46**
Activity/communityparticipation 0.02 0.16** 0.11* (cid:3)0.49** (cid:3)0.51**
Accesstoinformationandcommunication 0.12* 0.07 0.01 (cid:3)0.44** (cid:3)0.46**
Effortcompensatingforhearingloss(Q3m) 0.30** 0.08 (cid:3)0.03 (cid:3)0.34** (cid:3)0.35**
Stresswhentalkingonphone(Q3p) 0.21** 0.09 0.07 (cid:3)0.43** (cid:3)0.39**
Satisfactionlisteningtolectures(Q4g) 0.12* 0.07 0.07 (cid:3)0.33** (cid:3)0.49**
Satisfactionattendingmovies/plays(Q4h) 0.08 0.05 (cid:3)0.03 (cid:3)0.36** (cid:3)0.41**
Satisfactionlisteningtomusic(Q4i) 0.20** 0.09 0.01 (cid:3)0.44** (cid:3)0.34**
an¼376.
bn¼183.
Correlationswithoutasuperscriptwerenotsignificant(p40.05).
*p50.05.
**p(cid:4)0.01.
8 C. J.Stika&R. D.Hays
Multiple regression analysis andguidedbytheWHO-ICFtheoreticalframework.Inaddition,the
Multivariate ordinary least squares regression analyses were IHEAR-IT targets themes and domains identified by individuals
performed to determine whether the IHEAR-IT explained unique withAOHL as importantand relevant totheirlives.
variance in the participant’s global rating of quality of life and The 73-item IHEAR-IT was completed by a diverse sample of
appraisal of the impact of hearing loss on his/her quality of life. 409 individuals with varying degrees of AOHL, who resided in
Globalratingofoverallqualityoflifeandattributionoftheimpact different geographic areas of the United States. Multitrait scaling
ofhearinglossonqualityoflifewereeachassessedbyaone-item techniques provided support for four multi-item scales and five
index from the IHEAR-IT. The HHIE/HHIA social/situational and individualitems.FindingsindicatethattheIHEAR-ITdemonstrates
emotionalscalesweresignificantlyassociatedwithglobalratingsof excellent internal consistency reliability, with coefficient alpha
quality of life, F(2,309)¼39.16, p50.001, and explained 20% of easily exceeding the 0.70 minimum for group comparisons
theadjustedvariance.However,theadditionofsignificantIHEAR- (Nunnally & Bernstein, 1994). Support for the construct validity
IT scales (psychological/emotional and social/interpersonal), ofthefourscaleswasfoundbasedoncorrelationswiththeSF-36v2
resulted in 28% adjusted variance (8% additional unique variance MCSand theHHIE/HHIA.
explained).Thestrongestpredictoroftheglobalratingofqualityof Construct validity was furtherdemonstrated by confirmation of
lifewastheIHEAR-ITpsychological/emotionalscale(standardized hypothesizedpatternsofassociationsbetweentheIHEAR-ITscale
regression coefficients followed by zero-order correlations: scores and individual characteristics. Consistent with research on
(cid:3)¼0.33, p50.01; r¼0.52), followed by the IHEAR-IT social/ individuals with hearing loss (Hallberg et al, 2008; Helvik et al,
interpersonal scale((cid:3)¼0.30, p50.01;r¼0.48). The HHIE/HHIA 2006; Nachtegaal et al, 2009; Tambs, 2004), women and younger
social/situational scale ((cid:3)¼0.24, p50.05; r¼(cid:3)0.35) also had a individualsinourstudyreportedpoorerHRQOLontheIHEAR-IT
significant unique association with the global rating of quality of than did males and older individuals. Further, as anticipated, the
6
1 life, but the regression coefficient was a suppression effect (zero IHEAR-IT scales and its five individual items correlated signifi-
0
2 order correlation was negative but beta was positive). The HHIA/ cantlywithhearinglossseverity,suchthatgreaterhearinglosswas
ril HHIA emotional scale was not significantly, uniquely associated associated with poorer HRQOL. This finding is consistent with
p
A with the participants’ global rating of quality of life ((cid:3)¼(cid:3)0.14, previous research, including several large epidemiological studies
22 p¼0.17;r¼(cid:3)0.45). (e.g. Chia et al, 2007; Dalton et al, 2003), though correlations
9 The HHIE/HHIA scales were significantly associated with the reported have often been small. That the associations between
2
0: globalratingoftheimpactofhearinglossonoverallqualityoflife, hearing impairment and HRQOL found in this study are stronger
1
at F(2,309)¼110.32, p50.001, accounting for 41% of the adjusted thancorrelationsidentifiedinotherstudiesmaybeduetotheuseof
] variance. However, adding significant IHEAR-IT scales and items self-reportsofhearinglosslevelratherthanaudiometricmeasures.
a
n Stik IyHieEldAeRd-aInTadsojucsiatel/dinRte2ropfer5s0o%nal(9s%cauleniqaunedvtahreianHcHeIeEx/pHlaHinIeAd).emThoe- I(f20s0o4,)t,hiensewfhinicdhinsgeslfa-rreepionrtkedeehpeinagrinwgitlhostshoesxeplraeipnoerdtesdelbf-yreTpaomrtebds
e tional scale accounted for equal amounts of unique variance mental health and subjective well-being much better than did
r
r
a ((cid:3)¼0.30, p50.001; r¼0.64 and (cid:3)¼0.30, p50.01; r¼0.66, measured hearingloss.
C
y [ respectively); followed by the IHEAR-IT access to information Consistent with our expectations, the IHEAR-IT correlated
b scale ((cid:3)¼0.16, p50.05; r¼0.59); and finally, one IHEAR-IT significantly with the SF-36v2 MCS but not with the PCS;
d
e individual item (‘effort required to compensate for hearing loss’ however, the mean scores for the two scales were similar to
d
a [q3m] ((cid:3)¼0.18, p50.01; r¼0.53). The HHIA/HHIA social/ those for the U.S. general population. Studies that have used the
o
nl situational scale was not significantly, uniquely associated with SF-36v2 to evaluate HRQOL for individuals with hearing loss
w
o ratingof impact ofhearing lossonqualityof life(p¼0.11). have reported varied findings, with some showing an association
D
between greater hearing loss and poorer scores on both the SF-36
PCS and MCS (e.g. Chia et al, 2007; Dalton et al, 2003), other
Discussion
studies reporting no significant relationships (e.g. Chew & Yeak,
Thispaperdescribesthedevelopment,fieldtesting,andpreliminary 2010; Hickson et al, 2008; Hua et al, 2013; Parving et al, 2001),
psychometric evaluation of the IHEAR-IT, a HRQOL instrument and yet others, like ours, reporting significant associations only
for individuals with AOHL. Although other HRQOL measures for with the MCS (e.g. Abrams et al, 2002). Since hearing loss is
individualswithhearingimpairmentexist,theIHEAR-ITisunique considered a communication disability rather than a physical
in that it is a multidimensional HRQOL measure. Furthermore, disability or disease, it is not surprising that the SF-36v2 MCS
whiletheIHEAR-ITandtheHHIE/HHIAoverlapinsomecontent, would be more sensitive to the daily experiences and challenges
regression models demonstrated that the IHEAR-IT was signifi- for individuals with AOHL than would the PCS. On the other
cantly uniquely associated with the participants’ global rating of hand, a growing body of research is identifying significant
quality of life and appraisal of the impact of hearing loss on associations between hearing loss in older individuals and
everyday lifeactivities and well-beingbeyondtheHHIE/HHIA. cognitive decline, falls, and impaired activities of daily living
The development of the IHEAR-IT followed stringent proced- (Gopinath et al, 2012; Gurgel et al, 2014; Lin et al, 2013;
ures recommended by the HRQOL research community (Hays & Peracino, 2014). However, for the general population of individ-
Revicki, 2005; Reeve et al, 2013) and set forth by the FDA uals with hearing loss, especial younger individuals and individ-
GuidancetoIndustry(2009).Concertedeffortwasmadetoinclude uals who are otherwise physically healthy, the SF-36v2 and other
the perspective and advice of individuals with AOHL and various generic HRQOL measures, have been reported to lack sensitivity
professionalswhoservethem.Consistentwiththestudy’saim,the and specificity in assessing the impact of hearing loss on
IHEAR-IT provides a comprehensive assessment of the impact of HRQOL (Bess, 2000; Chew & Yeak, 2010; Mo et al, 2004).
AOHL, which is linked to current conceptual models of HRQOL Findings from this study support this conclusion, which