asystematic review of health-related quality of life and ... · j am acad audiol 18:151–184...

33
J Am Acad Audiol 18:151–184 (2007) 151 *Department of Communication Sciences and Disorders, University of South Florida, Tampa, FL; †Department of Communication Disorders, Auburn University, AL; ‡Department of Speech and Hearing Sciences, University of California Santa Barbara; §Bay Pines Veterans Affairs Healthcare System, Audiology and Speech Pathology Service, Bay Pines, FL; **University of Akron, School of Speech and Audiology, Akron, OH; ††Rush University Medical Center, Communication Disorders and Sciences, Chicago, IL 60612; ‡‡Cleveland Clinic, Head and Neck Institute, Cleveland, OH Carole E. Johnson, Ph.D., Au.D., Department of Communication Disorders, 1199 Haley Center, Auburn University, AL 36849; Phone: 334-844-9603; Fax: 334-844-4585; E-mail: [email protected] A Systematic Review of Health-Related Quality of Life and Hearing Aids: Final Report of the American Academy of Audiology Task Force on the Health-Related Quality of Life Benefits of Amplification in Adults Theresa Hnath Chisolm * Carole E. Johnson Jeffrey L. Danhauer Laural J.P. Portz * Harvey B. Abrams § Sharon Lesner ** Patricia A. McCarthy †† Craig W. Newman ‡‡ Abstract This is the final report of the American Academy of Audiology Task Force on the Health-Related Quality of Life (HRQoL) Benefits of Amplification in Adults. A systematic review with meta-analysis examined evidence pertaining to the use of hearing aids for improving HRQoL for adults with sensorineural hearing loss (SNHL). Relevant search strings applied to the C E N T R A L, CINAHL, Cochrane reviews, ComDisDome, EBMR, and PubMed databases identified randomized controlled trial, quasi-experimental, and nonexperimental pretest/posttest designed studies. Sixteen studies met a priori criteria for inclusion in this review. A random-effects meta-analysis showed differential results for generic versus disease-specific HRQoL measures for within- and between- subject designs. Although generic measures used for within-subject designs did not demonstrate HRQoL benefits from hearing aids, mean effect sizes and confidence intervals for within-subject designs and disease-specific instruments suggested that hearing aids have a small-to-medium impact on HRQoL. Further, the between-subject studies supported at least a small effect for generic measures, and when measured by disease-specific instruments, hearing aids had medium-to-large effects on adults’ HRQoL. This review concludes that hearing aids improve adults’ HRQoL by reducing psychological, social, and emotional effects of SNHL. Future studies should include control groups using randomized controlled trials. Key Words: American Academy of Audiology Task Force on the Health- Related Quality of Life Benefits of Amplification in Adults, health-related quality of life, hearing aids, hearing loss, meta-analysis, nonacoustic benefits, systematic review Abbreviations: AAA = American Academy of Audiology; ADPI-VAS = Auditory Disability Preference Index—Visual Analog Scale; CENTRAL = Cochrane Central Register of Controlled Trials; CI = confidence internal; CINAHL = Cumulative Index to Nursing and Allied-Health Literature; ComDisDome = Communication Sciences and Disorders DOME; EBM = evidence-based medicine; EBMR = evidence-based medicine review; EBP = evidence-based

Upload: others

Post on 14-Mar-2020

0 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: ASystematic Review of Health-Related Quality of Life and ... · J Am Acad Audiol 18:151–184 (2007) 151 *Department of Communication Sciences and Disorders, University of South Florida,

J Am Acad Audiol 18:151–184 (2007)

151

*Department of Communication Sciences and Disorders, University of South Florida, Tampa, FL; †Department ofCommunication Disorders, Auburn University, AL; ‡Department of Speech and Hearing Sciences, University of CaliforniaSanta Barbara; §Bay Pines Veterans Affairs Healthcare System, Audiology and Speech Pathology Service, Bay Pines, FL;**University of Akron, School of Speech and Audiology, Akron, OH; ††Rush University Medical Center, CommunicationDisorders and Sciences, Chicago, IL 60612; ‡‡Cleveland Clinic, Head and Neck Institute, Cleveland, OH

Carole E. Johnson, Ph.D., Au.D., Department of Communication Disorders, 1199 Haley Center, Auburn University, AL36849; Phone: 334-844-9603; Fax: 334-844-4585; E-mail: [email protected]

A Systematic Review of Health-Related Qualityof Life and Hearing Aids: Final Report of theAmerican Academy of Audiology Task Force onthe Health-Related Quality of Life Benefits ofAmplification in Adults

Theresa Hnath Chisolm*

Carole E. Johnson†

Jeffrey L. Danhauer‡

Laural J.P. Portz*

Harvey B. Abrams§

Sharon Lesner**

Patricia A. McCarthy††

Craig W. Newman‡‡

Abstract

This is the final report of the American Academy of Audiology Task Force onthe Health-Related Quality of Life (HRQoL) Benefits of Amplification in A d u l t s .A systematic review with meta-analysis examined evidence pertaining to theuse of hearing aids for improving HRQoL for adults with sensorineural hearingloss (SNHL). Relevant search strings applied to the C E N T R A L, CINAHL,Cochrane reviews, ComDisDome, EBMR, and PubMed databases identifiedrandomized controlled trial, quasi-experimental, and nonexperimentalpretest/posttest designed studies. Sixteen studies met a priori criteria forinclusion in this review. A r a n d o m - e ffects meta-analysis showed differential resultsfor generic versus disease-specific HRQoL measures for within- and between-subject designs. Although generic measures used for within-subject designsdid not demonstrate HRQoL benefits from hearing aids, mean effect sizes andconfidence intervals for within-subject designs and disease-specific instrumentssuggested that hearing aids have a small-to-medium impact on HRQoL.F u r t h e r, the between-subject studies supported at least a small effect forgeneric measures, and when measured by disease-specific instruments,hearing aids had medium-to-large effects on adults’ HRQoL. This reviewconcludes that hearing aids improve adults’ H R Q o L by reducing psychological,social, and emotional effects of SNHL. Future studies should include controlgroups using randomized controlled trials.

Key Words: American Academy of Audiology Task Force on the Health-Related Quality of Life Benefits of Amplification in Adults, health-related qualityof life, hearing aids, hearing loss, meta-analysis, nonacoustic benefits,systematic review

Abbreviations: A A A = American Academy of Audiology; A D P I - VAS = A u d i t o r yDisability Preference Index—Visual Analog Scale; CENTRAL = CochraneCentral Register of Controlled Trials; CI = confidence internal; CINAHL =Cumulative Index to Nursing and Allied-Health Literature; ComDisDome =Communication Sciences and Disorders DOME; EBM = evidence-basedmedicine; EBMR = evidence-based medicine review; EBP = evidence-based

Page 2: ASystematic Review of Health-Related Quality of Life and ... · J Am Acad Audiol 18:151–184 (2007) 151 *Department of Communication Sciences and Disorders, University of South Florida,

Journal of the American Academy of A u d i o l o g y/ Volume 18, Number 2, 2007

152

practice; EuroQoL = EuroQoL Group; EQ-5D = EuroQoL-5 dimensions; ES =e ffect size; EQ-VAS = E u r o Q o L - Visual Analog Scale; GDS = Geriatric DepressionS c a l e ; H H I A = Hearing Handicap Inventory for A d u l t s ; HHIE = Hearing HandicapInventory for the Elderly; H R Q o L = health-related quality of life; ICF =International Classification of Functioning, Disability, and Health; MOS SF-36= Medical Outcome Study 36-item Short-Form Health Survey; NIH = NationalInstitutes of Health; PTA = pure-tone average; QDS = Quantified Denver Scaleof Communicative Function; RCT = randomized controlled trial; SELF= Self-Evaluation of Life Function; SIGN = Scottish Intercollegiate Guideline Network;SNHL = sensorineural hearing loss; SPMSQ = Short Portable Mental StatusQ u e s t i o n n a i r e ; WHO = World Health Organization; WHO-DAS II = W o r l dHealth Organization-Disability Assessment Schedule

Sumario

Este es el reporte final del Grupo de Trabajo de la Academia Americana deAudiología en cuanto a los Beneficios en la Calidad de Vida Relacionados conla Salud (HRQoL) en la Amplificación en Adultos. Una revisión sistemática conmeta-análisis examinó la evidencia relacionado con el uso de auxiliares auditivos(AA) para mejorar la HRQoL de adultos con hipoacusia sensorineural (SNHL).Pistas relevantes de búsqueda aplicadas a las bases de datos del CINAHL, laBiblioteca Cochrane de Colaboración, el ComDisDome, la EBM y PubMedidentificaron estudios aleatorizados controlados, cuasi-experimentales, yestudios con diseño no experimental pre y post prueba. Dieciséis estudioscumplieron los criterios a priori para inclusión en esta revisión. Un meta-análisisde efectos aleatorios mostró resultados diferenciales para medidas de HRQoLgenéricas vs. específicas para enfermedad, en diseños inter-sujetos y intra-sujeto.Aunque las medidas genéricas usadas para los diseños intra-sujeto nodemostraron beneficios de los A A en HRQoL, los tamaños medios del efectoy los intervalos de confianza para los diseños intra-sujeto y los instrumentosespecíficos para enfermedad, sugieren que los A A tienen un impacto pequeñoa mediano en la HRQoL. Más aún, los estudios entre sujetos apoyaron al menosun pequeño efecto para las medidas genéricas, y cuando se midió parainstrumentos específicos de enfermedad, los A A tuvieron un efecto medianoa grande en la HRQoL de los adultos. Esta revisión concluye que los A Amejoran la HRQoL de los adultos, reduciendo los efectos psicológicos, socialesy emocionales de la SNHL. Los estudios futuros deberán incluir grupos de controlutilizando estudios aleatorizados controlados.

Palabras Clave: Grupo de Trabajo de la Academia Americana de Audiologíaen cuanto a los Beneficios de Calidad de Vida Relacionados con la Salud enla Amplificación en Adultos, calidad de vida relacionada con la salud, auxiliaresauditivos, hipoacusia, meta análisis, beneficios no acústicos, revisión sistemática

Abreviaturas: AAA = Academia Americana de Audiología; ADPI = Índice dePreferencia en Discapacidad Auditiva – Escala Visual Analógica; AHRQ =Agencia de Investigación y Calidad en Cuidados de Salud; CI = intervalo deconfianza; CINAHL = Índice Acumulativo en Literatura en Enfermería y Salud;ComDisDome = Ciencias y Trastornos de la Comunicación DOME; EBM =Medicina Basada en Evidencia; EBP = práctica basada en evidencia; EuroQoL= Grupo EuroQoL; EQ-5D = 5 dimensiones EuroQoL; ES = tamaño del efecto;EuroQoL EQ-VAS = EuroQoL – Escala visual Analógica; GDS = Escala deDepresión Geriátrica; HHIA = Inventario sobre Impedimento Auditivo paraAdultos; HHIE = Inventario sobre Impedimento Auditivo en el Viejo; HRQoL =calidad de vida relacionada con la salud; ICF = Clasificación Internacional deFunción, Discapacidad y Salud; MOS SF-36 = Forma Corta de la Encuestade Salud de 36 ítems sobre Resultados Médicos; NIH = Institutos Nacionalesde Salud; PTA = promedio tonal puro; QDS = Escala Cuantificada de Denversobre Función Comunicativa; RCT = estudio aleatorizado controlado; SELF= Auto-Evaluación de Función de Vida; SIGN = Red de Guía IntercolegialEscocesa; SNHL = hipoacusia sensorineural; SPMSQ = Cuestionario CortoPortátil sobre Estado Mental; WHO = Organización Mundial de la Salud;WHO – DAS II = Esquema de Evaluación sobre Discapacidad de la OMS

Page 3: ASystematic Review of Health-Related Quality of Life and ... · J Am Acad Audiol 18:151–184 (2007) 151 *Department of Communication Sciences and Disorders, University of South Florida,

Quality of Life and Hearing Aids/Chisolm et al

153

American Academy of Audiology Ta s kForce Members: Craig W. Newman, Co-chair;Theresa Hnath Chisolm, Co-chair; Harvey B.Abrams; Jeffrey L. Danhauer; Carole E.Johnson; Sharon Lesner; Patricia A .McCarthy; Laurel Portz.

This is the final report of the A m e r i c a nAcademy of Audiology (AAA) Ta s kForce on the Health-Related Quality of

Life Benefits of Amplification in Adults, whichwas created by A A A Past President A n g e l aLoavenbruck in 2003. The task force wascharged with conducting a systematic reviewof the evidence available pertaining to thenonacoustic benefits of amplification for adultswith sensorineural hearing loss (SNHL).

Hearing loss is a significant public healthproblem in the United States. It is the thirdmost common chronic health condition,exceeded only by arthritis and hypertension,in persons 65 years of age and older (HealthyPeople 2010, 2004). Although nearly 20million Americans over 45 years of age havea hearing loss, less than one-third of them usehearing aids (National Council on the A g i n g ,1999). The low proportion of adults withhearing loss who use hearing aids issurprising and may indicate that bothpatients with hearing loss and related health-care professionals are unaware of thepotential positive benefits that are availablefrom today’s hearing aids. It is especiallyimportant to document the benefits of hearingaids, because as the population continues toage and live longer, the number of peoplewith hearing loss will rise and cause anassociated increase in the demand for hearinghealth-care services.

Measuring the success of amplificationcan be accomplished by considering the Wo r l dHealth Organization’s (WHO) InternationalClassification of Functioning, Disability, andHealth (ICF; World Health Organization,2004). Within the ICF framework, the effectsof SNHL hearing loss and treatment outcomescan be examined at the levels of body functionsand structures, activity, and participation.Body functions are the physiological andpsychological tasks performed by the bodysystems (e.g., sensing the presence ordiscriminating the location, pitch, loudness,and quality of sounds), while body structuresare the anatomic parts of the body (e.g.,organs, limbs, and their components). Themost common type of hearing disorder inadults is SNHL, which primarily affects the

structures of the inner ear with the mostobvious consequences being a loss of hearingsensitivity or auditory impairment.

The negative consequences of adult-onsethearing loss are not limited to an auditoryimpairment; they can also involve activitylimitations and participation restrictions.According to the WHO-ICF, an activity is theexecution of a task or action by an individual,and participation refers to involvement inlife situations. Activity limitations arechanges at the level of the person (e.g.,inability to understand conversations), andparticipation restrictions are the effects ofthese limitations on broader aspects of life(e.g., withdrawing from social situations).C o n s e q u e n t l y, reductions in participationcan negatively impact an individual’s health-related quality of life (HRQoL).

S N H L is one of the few chronic conditionsfor which, in most cases, there are no effectivemedical or surgical treatments. However,audiologic intervention is available. Thecornerstone of the process of audiologicintervention is the use of amplification throughhearing aids, which is aimed at reducing theauditory impairment and optimizing thei n d i v i d u a l ’s auditory activities and minimizingparticipation restrictions (Kiessling et al, 2003).

Using the ICF framework, the benefits ofamplification can be assessed by examiningreductions in impairments, activity limitations,and/or participation restrictions (Abrams andHnath Chisolm, 2000). Both impairment leveland acoustic outcomes can be quantified bydocumenting hearing aid users’ aided overunaided improvements in audibility andspeech recognition using objective audiologictests. Acoustic outcomes may translate intoreductions in activity limitations that can bedocumented through the use of subjective self-report measures that assess auditoryperformance in a variety of listening situations.

At the level of participation, the benefits of amplification are related toreductions in any of the myriad ofpsychosocial problems associated withuntreated SNHL. Documented problemsinclude social isolation, depression, anxiety,and loneliness; lessened self-efficacy andmastery; and stress in relationships whenf a m i l y, friends, and coworkers experiencefrustration, impatience, anger, pity, and/orguilt while interacting with a person who hasa hearing loss (Weinstein and Ve n t r y, 1982;Bess et al, 1989; Uhlmann et al, 1989;Andersson and Green, 1995; Campbell et al,

Page 4: ASystematic Review of Health-Related Quality of Life and ... · J Am Acad Audiol 18:151–184 (2007) 151 *Department of Communication Sciences and Disorders, University of South Florida,

1999; Keller et al, 1999; National Council onthe Aging, 1999; Wayner and A b r a h a m s o n ,2001; Kramer et al, 2002). It is well knownthat mental, emotional, and socialconsequences of untreated hearing loss inadults can have negative impacts oni n d i v i d u a l s ’ overall HRQoL (Bess et al, 1989;Bess et al, 1990; Mulrow et al, 1990; Kelleret al, 1999; National Council on the A g i n g ,1999; Strawbridge et al, 2000; Dalton et al,2003; Pugh, 2004). Thus, assessing the effectsof amplification in terms of changes in HRQoLinvolves examination of myriad nonacousticb e n e f i t s .

H R Q o L assessments examine the degreeto which people’s health status affects theirself-perception of daily functioning and well-being. Quantitative measurements of HRQoLcan be made using generic and/or disease-specific instruments (National Institutes ofHealth [NIH], 1993). Generic instrumentsare broad in scope and applicability, whiledisease-specific instruments focus on onecondition, attempting to define its effects ondaily functioning and well-being. Evaluatingthe benefits of hearing aid use with disease-specific HRQoL instruments is appealingfrom a clinical perspective, because they arehighly responsive to interventions designedto manage a particular disease or disorder(Deyo and Patrick, 1989). Meanwhile,evaluating the benefits of hearing aids withgeneric HRQoL instruments is a timelye n d e a v o r, because there is increasedemphasis on their use across a broad rangeof health-related disciplines. Interest inH R Q o L measurement arises from severalfactors including (1) a shift in the focus fromlife prolongation to maintenance of anadequate HRQoL as one ages (i.e., living well;not merely living longer), (2) a generalagreement about the importance of patients’self-perceptions of health, and (3) the use ofH R Q o L measures to conduct health statuscomparisons across different conditions and/ortarget populations (Ware and Sherbourne,1992; Ware et al, 1993; McCallum, 1995;B r u l e y, 1999). Moreover, in this era of evidence-based practice (EBP), increased accountability,and shrinking health-care resources, it isreasonable for administrators, third-partypayers, and adults with hearing impairmentand their families to ask how different aspectsof audiologic treatment impact HRQoL(Abrams and Hnath Chisolm, 2000).

Within the EBP framework, thesystematic and explicit evaluation and

synthesis of a body of research evidence aboutspecific diagnostic procedures, medicaltreatments, and/or rehabilitativeinterventions is referred to as a “systematicr e v i e w.” According to the CochraneCollaboration (Higgins and Green, 2005)—which is an international nonprofit andindependent organization, dedicated toproducing and disseminating up-to-date andaccurate systematic reviews of evidence fromclinical trials and other studies about theeffects of health-care interventions, and tomaking their findings readily availableworldwide—the systematic review processinvolves:

• assessing amassed evidence abouttargeted interventions from afoundation of clearly formulatedquestions;

• systematically using explicit methodsto identify, select, critically appraise,and narrow down a vast body ofliterature to focus in on only themost relevant research pertainingto the topic; and

• collecting, analyzing, aggregating,and interpreting data from pertinentstudies to answer the question.

Thus, systematic reviews differ fromtraditional literature reviews by strictlyadhering to scientific design principles, whichshould result in reviews that are morecomprehensive, less likely to be biased, andmore apt to be reliable. All systematic reviewsallow for a qualitative summary of the resultsof individual studies to be completed.H o w e v e r, the Cochrane Collaboration(Higgins and Green, 2005) and otherorganizations that promote the use of EBP,including the Agency for Health Care Qualityand Research (AHRQ; www.ahrq.gov) andthe Scottish Intercollegiate GuidelineNetwork (SIGN; www.sign.ac.uk), alsoencourage the development of quantitativesystematic reviews. In a quantitativesystematic review, a statistical procedureknown as a “meta-analysis” is completed,which allows for the integration of the resultsof several independent studies. Bothqualitative and quantitative systematicreviews can provide more objective appraisalsof the evidence on a given topic thantraditional reviews. The inclusion of a meta-analysis in a systematic review allows formore precise estimates of treatment effects.Although the importance of EBP and the

Journal of the American Academy of A u d i o l o g y/ Volume 18, Number 2, 2007

154

Page 5: ASystematic Review of Health-Related Quality of Life and ... · J Am Acad Audiol 18:151–184 (2007) 151 *Department of Communication Sciences and Disorders, University of South Florida,

Quality of Life and Hearing Aids/Chisolm et al

155

systematic review process to audiologyreceived attention in a 2005 special issue ofthe Journal of the American Academy ofA u d i o l o g y, only a few quantitative systematicreviews have been completed, particularlyregarding the assessment of the effectivenessof amplification.

The charge of the present A A A task forcewas to examine nonacoustic benefits ofhearing aids. To accomplish this goal, thetask force members chose to conduct aquantitative systematic review of the evidenceassessing the effects of the use ofamplification by individuals with SNHL ontheir HRQoL compared to their unaidedcondition. The specific objective of thisinvestigation was to determine if the use ofhearing aids as compared to not using hearingaids resulted in improvements in HRQoL f o rindividuals with SNHL as measured throughboth disease-specific and generic instruments.

METHODS

When conducting systematic reviews, itis important for the reviewers to

establish at the outset the criteria by whichstudies will be included or excluded fromconsideration. The reviewers should pursuetheir search for relevant articles in the samefashion that a well-controlled research studywould be conducted (i.e., the reviewersthemselves become the investigators).Customarily, a team of investigators agreeson a number of a priori criteria that must besatisfied by the studies in order for them tobe included in a systematic review. Forexample, the investigators must agree onthe types of research designs, participants’a g eand severity of hearing loss, interventionstrategies, and outcome measures. The studyselection criteria for the present systematicreview are described below.

Study Selection Criteria

Types of Studies

The highest priority was given to studiesusing a randomized controlled trial (RCT)design. Double-blinded RCTs are the designof choice for examining the efficacy oftreatments, because nonrandomized,nonblinded studies may exaggerate the effects

of health-care interventions by up to 40%(Schulz et al, 1995). However, if only RCTs areincluded in a systematic review, thenimportant and relevant information may belost when comparing participants’ data fromunaided to aided conditions for at least tworeasons. First, participants cannot be blindedto the treatment arm, because there is nopractical way of providing a “placebo” whencomparing hearing aids as an intervention tono intervention. Second, using randomassignment to a delayed treatment conditionas a method of control may prolongparticipants from receiving hearing aidintervention, which could negatively impacttheir HRQoL. Therefore, to limit theinvestigation to only RCTs, to the exclusion ofother experimental designs, could preclude therecovery of important evidence demonstratingenhanced HRQoL from hearing aids.

A convenient way of summarizing studies’designs is to assign them a level of evidence.Level of evidence refers to the ability of aparticular study design to minimize oreliminate bias in the effect being measured.Several different categorization schemes areavailable for classifying a body of literaturein terms of a hierarchy of study design (AHRQ,2002), and reviewers must determine themost relevant level of evidence hierarchy forthe type of health-care procedure under review( R o b e y, 2004). The hierarchy of evidence usedin this study was adapted from the ScottishIntercollegiate Guideline Network (SIGN)system. The SIGN rating system is moredetailed than the adapted version used herein that it also assigns pluses and minuses todesignate further subcategories within thefirst two levels of evidence, a stratification thatwas not deemed to be necessary for thisr e v i e w. Studies accepted for inclusion in thisreview involved: Level 1—RCTs; Level 2—quasi-experimental controlled trials that usednonrandomized, parallel group, or crossoverdesigns; and Level 3—well-designednonexperimental studies, particularly thoseusing pretest/posttest designs with adequatedescriptions. Although studies involving lessstringent levels of evidence involving patienttestimonials or expert opinions on (Level4)were used for might provide valuable insightabout the HRQoL benefits of hearing aids,they were not considered to be of sufficientrigor for establishing EBP. Table 1 shows thehierarchy of evidence used in this s y s t e m a t i creview.

Page 6: ASystematic Review of Health-Related Quality of Life and ... · J Am Acad Audiol 18:151–184 (2007) 151 *Department of Communication Sciences and Disorders, University of South Florida,

Types of Participants

In order for studies to meet the inclusioncriteria for this review, their participantshad to be new or previous hearing aid usersof at least 18 years of age with normalcognitive function, independent or assistedliving conditions, and SNHL with unaidedseverity ranging from mild to profound.

Types of Interventions

This review examined the HRQoL b e n e f i t sof hearing aids; however, it was not designedto compare various technical aspects of theevaluation, selection, and fitting of amplification.Thus, all studies meeting the inclusion criteriawere considered regardless of the type of hearingaid style (e.g., behind-the-ear, in-the-ear,completely-in-the ear, etc.), signal processingcircuitry (e.g., analog or digital), microphonetechnology (e.g., omnidirectional or directional),or fitting strategy (e.g., monaural vs. binaural)they employed.

Types of Outcome Measures

Another criterion for studies to beincluded in this systematic review was thatthey employed only previously validatedgeneric and/or disease-specific HRQoLinstruments as outcome measures. Recallthat HRQoL instruments measure the degreeto which participants’ health status affectstheir self-perception of daily functioning andwell-being. Generic instruments areapplicable across diseases and disorders,while disease-specific instruments aredesigned for use with a specific patient

population. Studies accepted for inclusion inthis review employed widely used genericoutcome measures, such as the M e d i c a lOutcome Study 36-item Short-Form HealthSurvey (MOS SF-36; Ware and Sherbourne,1992) and the E u r o Q o L - 5 d i m e n s i o n s ( E Q -50; EuroQoL Group, 1990). In addition, thesystematic review included studies employingdisease-specific self-report instruments, suchas the Hearing Handicap Inventory for theE l d e r l y (HHIE; Ventry and Weinstein, 1982),which measure the effects of hearing loss inthe psychological, social, and emotionaldomains.

Search and Retrieval Process for theIdentification of Studies

A full search strategy was developed toidentify studies to include in the systematicr e v i e w. Six of the eight authors were dividedinto three two-person teams with one of theremaining authors (T.C.) serving as amediator for differences of opinion in thesearch and retrieval process, and the other( L . P.) serving as the compiler of search results.One team (H.A. and P.M.) searched non-peer-reviewed journals (e.g., The Hearing Journaland Hearing Review) and recent professionalconference proceedings to obtain potentiallyrelevant studies. The other two teams (C.J.and J.D., Team 1; and C.N. and S.L., Team2) conducted database searches of peer-reviewed journals. The reviewers createdsearch strings using terms that were mostlikely to result in the greatest number ofapplicable hits from the database searches.Search strings that were used are shown inTable 2.

Journal of the American Academy of A u d i o l o g y/ Volume 18, Number 2, 2007

156

Table 1. Levels of Evidence Used for Rating Studies in This Review as Adapted from the ScottishIntercollegiate Guideline Network (SIGN) System

1 High quality meta-analyses, systematic reviews of RCTs, or RCTs with a very low risk of bias

1 Well-conducted meta-analyses, systematic reviews of RCTs, or RCTs with a low risk of bias

1 Meta-analyses, systematic reviews of RCTs, or RCTs with a high risk of bias

2 High-quality systematic reviews of case-control or cohort studiesHigh-quality case-control or cohort studies with a very low risk of confounding, bias, or chance, and a high probability that the relationship is causal

2 Well-conducted case-control or cohort studies with a low risk of confounding, bias, or chance, and a moderate probability that the relationship is causal

2 Case-control or cohort studies with a high risk of confounding, bias, or chance, and a significant risk that the relationship is not causal

3 Nonanalytic studies (e.g., case reports, case series)

4 Expert opinion

Page 7: ASystematic Review of Health-Related Quality of Life and ... · J Am Acad Audiol 18:151–184 (2007) 151 *Department of Communication Sciences and Disorders, University of South Florida,

Quality of Life and Hearing Aids/Chisolm et al

157

The databases included theCommunication Sciences and DisordersDOME (ComDisDome) searched by Team 1,as well as the Cumulative Index to Nursingand Allied-Health Literature (CINAHL),Evidence-Based Medicine Reviews (EBMR);the Cochrane Central Register of ControlledTrials and the Cochrane Database ofSystematic Reviews (Cochrane Reviews)searched by Team 2. The ComDisDome tapsinto PubMed, select journals from the

Linguistics and Language Behavior A b s t r a c t sfor non-PubMed journals, Seminars inH e a r i n g, select awarded grants from theNational Institute of Health (NIH) and theNational Science Foundation (NSF),dissertations from ProQuest, books fromcustom-selected multiple publisher sources,author profiles drawn from its ownproprietary Scholar Universe database, andcustom Web sites. Team 1’s search includedarticles published in peer-reviewed journalssince 1980; Team 2’s search included thosepublished since 1996. The search was limitedto articles published in English, and thedatabase search occurred during July 2004.

RESULTS

Study Flow

Figure 1 summarizes the search and

Figure 1. Flow chart for the search and retrievalprocess for articles to include in the systematic review.

Table 2. Search Strings Used in the Databases• Abbreviated Profile of Hearing Aid Benefit (APHAB)• Amplification and Quality of Life in Adults• Amplification, Hearing Aids, and Depression• Amplification, Hearing Aids, and Functional Health

Status• Amplification, Hearing Aids, and Health Status• Amplification, Hearing Aids, and Interpersonal

Relationships• Amplification, Hearing Aids, and Isolation• Amplification, Hearing Aids, and Mental Status• Amplification, Hearing Aids, and Paranoia• Amplification, Hearing Aids, and Participation• Amplification, Hearing Aids, and Psychosocial Function• Amplification, Hearing Aids, and Sadness• Amplification, Hearing Aids, and Self-Concept• Amplification, Hearing Aids, and Social Activity• Amplification, Hearing Aids, and Well-Being• Amplification, Hearing Aids, and Withdrawal• Benefit of Amplification in Adults• Client Oriented Scale of Improvement• Communication Profile for the Hearing Impaired• Denver Scale of Communication Function• Effects of Hearing Aid Use in Adults• Glasgow Hearing Aid Benefit Profile• Hearing Aid Benefit in Adults• Hearing Aid Performance Inventory• Hearing Aids and Quality of Life in Adults• Hearing Handicap Inventory for Adults• Hearing Handicap Inventory for the Elderly• Hearing Handicap Scale• Hearing Measurement Scale• Hearing Performance Inventory• International Outcome Inventory• Medical Outcomes Study Short Form 36 and Amplification• Medical Outcomes Study Short Form 36 and Hearing• Medical Outcomes Study Short Form 36 and Hearing Aids• Nonacoustic Benefits of Amplification in Adults• Nonacoustic Benefits of Hearing Aids in Adults• Performance Inventory for Profound and Severe Loss• Profile of Hearing Aid Benefit• Self-Assessment of Communication • Shortened Hearing Aid Performance Inventory• Sickness Impact Profile and Hearing• Sickness Impact Profile and Hearing Aids• Sickness Impact Profile and Amplification• Significant Other Assessment of Communication• Subjective Measures and Amplification• Subjective Measures and Hearing Aids• Weighted Index of Social Hearing Handicap

Page 8: ASystematic Review of Health-Related Quality of Life and ... · J Am Acad Audiol 18:151–184 (2007) 151 *Department of Communication Sciences and Disorders, University of South Florida,

retrieval process used in this systematicr e v i e w. Using the key words and databasesdescribed above, the search and retrievalprocess identified 171 relevant abstracts,excluding duplicates. Reviewing the abstractsof all 171 studies revealed that 96 did notmeet inclusion criteria, and thus, they were notconsidered for further review. Recall thatarticles for inclusion in this systematic reviewhad to involve adult participants with SNHLand use outcome measures assessing HRQoL,and had to be at an appropriate level ofevidence. Because a review of the abstracts didnot fully reveal whether the remaining 75studies met inclusion criteria, a full articlereview was conducted on all 75 of them. Thesestudies were randomly distributed to the threeteams for consideration and categorizationinto included and excluded studies for thesystematic review. Team membersindependently judged the articles with 90% orbetter interjudge agreement. A third judge( T.C.) settled any disagreements throughmediation resulting in a group consensus.

This process yielded 16 articles forinclusion; 58 were excluded from furtherconsideration in this investigation. Note,h o w e v e r, that one (Yueh et al, 2001) of the 16articles that was categorized as included wasexcluded from the quantitative analysisbecause the methods used for reporting theresults precluded extraction of the data foruse in the meta-analysis conducted as partof this systematic review. However, data frompart of the study were suitable for inclusionin the qualitative analyses, and thus, theyueh et al (2001) study was listed as one ofthe 16 included studies. The journal citingsfor the excluded articles are provided inAppendix A, and Appendix B indicates thereason(s) why each of these studies wasexcluded from this review.

Study Characteristics

Appendix C provides information aboutthe design methodologies, participantcharacteristics, and outcome measures usedby the 16 studies that were included in thissystematic review.

Design Methodology

Only two (Mulrow et al, 1990; Yueh et al,2001) of the 16 studies used RCTs, a design

of the highest level of evidence. The overalldesigns used in five (Abrams et al, 1992;Newman et al, 1993; Chmiel and Jerger,1996; Jerger et al, 1996; Primeau, 1997) of thestudies were quasi-experimental. Of those,only Abrams and colleagues (1992) used acontrol group. However, their participantswere not randomly assigned to the control andthe hearing aid intervention groups. A l t h o u g hthe other four of these studies were quasi-experimental, they addressed questions thatwere not relevant to the present review and,thus, were grouped with the studies thatused the pretest/posttest design. For example,Jerger and colleagues (1996) used a crossoverdesign, in which all participants receivedeach of three treatments, to compareoutcomes with conventional hearing aids,assistive listening devices, and bothinterventions combined. Newman et al (1993),Chmiel and Jerger (1996), and Primeau(1997) used a between-groups comparison inwhich both groups were fitted with hearingaids and the effects of intervention wereexamined as a function of group membership.The most common methodology was thenonexperimental pre-post test design, whichwas used in 13 (the remaining nine plus thefour just mentioned) of the 16 studies. Forexample, Dillon et al (1997) used this designto examine relationships among a variety ofoutcome measures.

Participant Characteristics

Participant characteristics of age, gender,hearing aid experience, and degree of hearingloss were examined. The hearing aid deliverysystem used with the participants in eachstudy was also noted.

AgeAll but one of the 16 studies used

participants whose mean age was 60 years oro l d e r. The only exception (Primeau, 1997)compared outcomes for groups of youngerand older individuals. The ranges andstandard deviations of ages in most of thestudies were large and suggested theinclusion of young and middle-aged adults inthe samples. For example, Joore et al (2002)and Joore et al (2003) cited a lower age limitof 28 years, while Jerger et al (1996) andPrimeau (1997) reported age limits exceeding95 years of age. Thus, there was considerable

Journal of the American Academy of A u d i o l o g y/ Volume 18, Number 2, 2007

158

Page 9: ASystematic Review of Health-Related Quality of Life and ... · J Am Acad Audiol 18:151–184 (2007) 151 *Department of Communication Sciences and Disorders, University of South Florida,

Quality of Life and Hearing Aids/Chisolm et al

159

heterogeneity among the participants’ agesacross the studies.

GenderOnly one study (Dillon et al, 1997) failed

to report the gender distribution of itsparticipants. Of the remaining 15 studies, 11used both male and female participants,while three (Newman and Weinstein, 1988;Abrams et al, 1992; Yueh et al, 2001) usedonly male participants from Ve t e r a n sAdministration Medical Centers.

Degree of Hearing LossThe participants’ hearing levels revealed

considerable diversity in how these data werecollected and presented. For example, mostof the studies reported pure-tone averages( P TAs), but the frequencies used in theircalculation varied. In addition, some studiesreported PTAs for each ear while others onlypresented them for the ear fitted with ahearing aid, or for the better ear in the caseof binaural hearing aid fittings. A few trendsemerged despite this variability. For example,although all of the studies reported meanPTAs in the mild-to-moderate range, theirstandard deviations and ranges revealed thatthey also included at least some participantswhose hearing losses were severe for at leastsome frequencies. Also, none of the studiesappeared to have included participants withPTAs in the profound range.

Hearing Aid ExperienceOf the 16 studies, 15 involved new

hearing aid users as participants. Of these 15studies, three included experienced as well asnew hearing aid users (i.e., Jerger et al, 1996;Primeau, 1997; Humes et al, 2001). Newmanand Weinstein (1988) only includedexperienced hearing aid users.

Hearing Aid Delivery SystemThe participants in six of the studies

(Newman and Weinstein, 1988; Mulrow et al,1990, 1992a; Abrams et al, 1992; Primeau,1997; Yueh et al, 2001) received hearing aidsthrough the Veterans A d m i n i s t r a t i o nNational Hearing Aid Program. Four of thestudies (Dillon et al, 1997; Joore et al, 2002,2003; Stark and Hickson, 2004) wereconducted in countries where hearing aidswere provided through national healthprograms. In the studies by Chmiel and Jerger(1996) and Jerger et al (1996), data were

obtained as part of a trial period of hearingaid use. In one study (Newman et al, 1993),half of the participants received hearing aidsas part of their medical insurance benefits,and the other half were private pay. In two ofthe remaining studies (Ta y l o r, 1993; Humeset al, 2001), participants paid out-of-pocket fortheir hearing aids. Only Malinoff andWeinstein (1989a) did not report how theirp a r t i c i p a n t s ’ hearing aids were funded.

Outcome Measures

Nine different outcome measures wereused across the 16 studies. Eleven of thestudies used a single outcome measure.Several of the studies used self-reportmeasures, which were not considered to assessH R Q o L in this review. The remaining fivestudies used multiple HRQoL i n s t r u m e n t s .Five generic and four disease-specificinstruments were used in these studies. Thefive generic tools were the E Q - 5 6 ( E u r o Q o LGroup, 1990), the Geriatric Depression Scale(GDS; Yesavage et al, 1982–83), the MOS SF-36 (Ware and Sherbourne, 1992), the S e l f -Evaluation of Life Function (SELF; Linn andLinn, 1984), and the Short Portable MentalStatus Questionnaire (SPMSQ; Pfeiffer, 1975).The four disease-specific outcome measuresincluded the Auditory Disability PreferenceI n d e x — Visual Analog Scale ( A D P I - VAS; Jooreet al, 2002), the Hearing Handicap Inventoryfor Adults (HHIA: Newman et al, 1990), theHearing Handicap Inventory for the Elderly(HHIE; Ventry and Weinstein, 1982), and theQuantified Denver Scale of CommunicationF u n c t i o n (QDS; Tuley et al, 1990).

The E Q - 5 D, used in two studies (Joore etal, 2002, 2003), is a two-part, self-reportquestionnaire. The first part assesses fivedimensions of HRQoL (i.e., mobility, self-care, usual activities, pain or discomfort, andanxiety or depression) on a three-point ordinalscale. The second part is a 20 cm visualanalog scale EQ-VAS ranging from 0 (worstimaginable health state) to 100 (bestimaginable health state) on which patientsestimate their current state of health.

Mulrow et al (1990, 1992a) used the GDS,SELF, and SPMSQ. The GDS is a 15-itemscale scored from 0 to 15 and is commonlyused as a screening tool for assessingdepression in older individuals. The SPMSQis a 10-item scale with scores ranging from

Page 10: ASystematic Review of Health-Related Quality of Life and ... · J Am Acad Audiol 18:151–184 (2007) 151 *Department of Communication Sciences and Disorders, University of South Florida,

1 to 10 for assessing cognitive function. TheSELF is a 54-item global scale assessing sixdomains (i.e., physical disability, socialsatisfaction, symptoms of aging, depression,self-esteem, and personal control) andprovides scores ranging from 54 to 216.Higher scores indicate greater dysfunction onthe SELF, GDS, and SPMSQ.

The MOS SF-36, used in two studies(Joore et al, 2003; Stark and Hickson, 2004),is a 36-item questionnaire providing scoresin eight domains (i.e., physical function, role-function physical, bodily pain, general health,v i t a l i t y, social function, role-functionemotional, and mental health) with higherscores indicating better functioning. Notethat Joore et al (2003) reported on only theMOS SF-36 social function subscale. Rawscores or standardized T-scores can bereported for each of the MOS SF-36 subscales,with lower scores indicating greaterdysfunction. Although the subscales of theMOS SF-36 also combine to provide physicaland mental component summary scores, theywere not reported in the studies reviewed.

The HHIE was the most commonly useddisease-specific instrument. The HHIE is a25-item questionnaire that assesses theimpact of hearing loss. It provides scoresranging from 0 to 48 and 0 to 52 for social andemotional subscales, respectively, andcombined scores ranging from 0 to 100. TheHHIA, used in one study (Primeau, 1997), isnearly identical to the HHIE but has a fewitems that were modified for patients under65 years of age. Higher scores on both theHHIE and HHIA scales indicate greaterperceived hearing handicap. One study(Dillon et al, 1997) modified the HHIE byreversing the scale so that 0 indicated greaterhandicap and 4 indicated lesser handicap.Thus, rather than resulting in a score from0 to 100, their group data were presented ona scale of 0 to 4. In addition, the QDS is a 25-item questionnaire, which yields scoresranging from 0 to 100, with higher scoresindicating greater difficulty. The QDS wasused in two studies (Mulrow et al, 1990,1992a), and it assesses both communicationand emotional difficulties resulting fromhearing loss. Finally, the A D P I - VAS was usedin two studies (Joore et al, 2002, 2003). It alsoresults in scores ranging from 0 to 100, withhigher scores indicating better functioning.

Study Quality

The quality of the evidence provided byeach of the 16 studies was examined byassessing the methods that they used tominimize bias and to control for potentiallyconfounding variables. The level of evidenceassigned to each of the 16 studies is shownin column 2 of Table 3.

Other study quality indicators assessedwere use of experimental and control groups,equivalence of experimental and controlgroups at baseline, a power analysis to ensureappropriate sample size, adequate detail ofparticipant inclusion and exclusion criteria,well-described hearing aid fitting andverification protocols, application andreporting of statistical analyses, andaccounting for any dropouts of participantsfrom the studies.

Table 3 also shows the evidence levelsassigned to each study. Most of the studiesprovided evidence at Level 3, with only one atLevel 2 (Abrams et al, 1992) and two at Level1 (Mulrow et al, 1990; Yueh et al, 2001). Thus,the quality indicator equivalence at baselinewas only applicable in the Abrams et al (1992),Mulrow et al (1990), and Yueh et al (2001)studies. All three of those studies reportedthat their control and treatment groups wereequivalent at baseline for age and hearingloss. Although additional studies used two ormore participant groups (e.g., Stark andHickson, 2004), they did not differ on factorsthat were relevant for the present review.Only one of the studies (i.e., Stark and Hickson,2004) reported use of a power analysis.

Table 3 shows that all but four of the 16studies provided sufficiently detailedparticipant inclusion and exclusion criteriato permit replication and generalization oftheir results to a specific population. Onlynine of the studies provided sufficient detailabout the hearing aid fittings, and all but oneof those studies provided information aboutthe verification of the fittings. Recall thatthis systematic review examined the HRQoLbenefits of hearing aids without considerationof the type of fitting or verification procedures.Thus, the studies that did not provide detailedinformation were still included in the review.In addition, all 16 of the studies were judgedto have used appropriate statistical analyses.H o w e v e r, only 10 of the studies mentioned thenumber of participants who did not completethe study (i.e., “dropouts”), and only five of

Journal of the American Academy of A u d i o l o g y/ Volume 18, Number 2, 2007

160

Page 11: ASystematic Review of Health-Related Quality of Life and ... · J Am Acad Audiol 18:151–184 (2007) 151 *Department of Communication Sciences and Disorders, University of South Florida,

Quality of Life and Hearing Aids/Chisolm et al

161

those provided sufficient description ofreasons for their participants’ failure tocomplete the protocols.

Qualitative Assessment of StudyResults

Effects of Hearing Aid Use on HRQoLas Measured through GenericInstruments

Comparing pre- and posttest results fordimensions or scales on some of the genericH R Q o L measures showed significantlyimproved health states of participantsfollowing hearing aid intervention. Forexample, Joore et al (2002, 2003)demonstrated reduced anxiety and depression

as measured by the EQ-5D for new hearingaid wearers. This finding is consistent withthose of Mulrow and colleagues (1990), whofound significantly reduced states ofdepression as measured by the GDS in agroup of adults with SNHL who wereimmediately treated with hearing aids ascompared to a control group of peers. Mulrowand colleagues (1990, 1992a) also foundpositive improvements in mental functioningon the SPMSQ in the same treatment groupafter hearing aid use as compared to thecontrol group. A d d i t i o n a l l y, Joore et al (2003)demonstrated greater social functioningpost–hearing aid fitting as compared top a t i e n t s ’ pretreatment states on the MOSSF-36. However, Stark and Hickson (2004)showed a reduction in social functioning andvitality after their participants were fit withhearing aids compared to their pretreatment

Table 3. Quality Assessment of the 16 Studies Included in the Qualitative Analysis for This Systematic ReviewStudy Level of Control Baseline Power Inclusion HA Verification Appropriate Drop-outs

Evidence Group Equivalence Analysis Exclusion Fit of HA Fit Statistics Discussed

Abrams et al (1992) 2 Y Y N Y N N Y NA

Chmiel and Jerger (1996) (2) 3 N N/A N Y Y Y Y NA

Dillon et al (1997) (2) 3 N/A N/A N N Y Y Y NA

Humes et al (2001) 3 N N/A N Y Y Y Y NA

Jerger et al (1996) (2) 3 N N/A N Y Y Y Y NA

Joore et al (2002) 3 N N/A N Y N Y Y Y

Joore et al (2003) 3 N N/A N Y Y Y Y Y

Malinoff and Weinstein (1989a) 3 N N/A N N N N Y NA

Mulrow et al (1990) 1 Y Y N Y N N Y Y

Mulrow et al (1992a) 3 N N/A N Y N N Y Y

Newman et al (1993) 3 N N/A N N Y Y Y N

Newman and Weinstein (1988) 3 N N/A N Y N N Y Y

Primeau (1997) 3 N N/A N N N N Y N

Stark and Hickson (2004) 3 N Y Y Y Y Y Y N

Taylor (1993) 3 N N/A N Y Y N Y N

Yueh et al (2001) 1 Y Y N Y Y Y Y NThis study was usedin the qualitative analysis, but not thequantitative analysis.

Note: Level 1 = randomized controlled trials; Level 2 = quasi-experimental; Level 3 = nonexperimental. When two study levels are shown, the oneoutside of the parenthesis was used for the analyses completed in this review. Y = yes; N = no; N/A = not applicable; HA = hearing aid.

Page 12: ASystematic Review of Health-Related Quality of Life and ... · J Am Acad Audiol 18:151–184 (2007) 151 *Department of Communication Sciences and Disorders, University of South Florida,

states. Also, Joore et al (2002, 2003) found nosignificant differences on any of the EQ-5Dscales of mobility, self-care, daily activities,or pain/complaints for their participants’ p r e -and post-hearing-aid scores. In addition,Stark and Hickson (2004) found no significantpre- and post-hearing-aid differences in agroup of participants with SNHL as measuredby the physical function, role-functionphysical, bodily pain, general health, role-function emotional role, or mental healthscales of the MOS SF-36.

Effects of Hearing Aid Use on HRQoLas Measured through Disease-SpecificInstruments

Contrary to the findings for the genericinstruments, most of the results for thedisease-specific HRQoL outcome measuresshowed strong reductions in the emotionaland social impacts of hearing loss forparticipants as a result of hearing aid usewhen measured by the HHIE (Newman andWeinstein, 1988; Malinoff and We i n s t e i n ,1989a; Mulrow et al, 1990, 1992a; Abrams etal, 1992; Newman et al, 1993; Taylor, 1993;Chmiel and Jerger, 1996; Jerger et al, 1996;Dillon et al, 1997; Primeau, 1997; Humes etal, 2001; Yueh et al, 2001; Stark and Hickson,2004) and the HHIA (Primeau, 1997).A d d i t i o n a l l y, similar significant reductions inparticipation restriction were obtained onthe QDS (Mulrow et al, 1990, 1992a).

Quantitative Assessment (Meta-analysis) of Study Results

Thoughtful, systematic qualitativereviews of research results provide importantinformation. However, conclusions based onqualitative analysis alone cannot substitutefor scientific evidence that is gatheredthrough the application of hypotheses-testinglogic using meta-analysis. Meta-analysis isthe mathematical synthesis of theindependent research findings from studiespublished throughout the literature (Robeyand Schultz, 1993). The inclusion of studiesfor the purposes of meta-analysis isdependent on their reporting of informationthat is needed to compute an effect size (ES),which is a metric that expresses the overallmagnitude of a result. In the present context,all of the studies included in the qualitativeanalysis, except Yueh et al (2001), providedsufficient information for the calculation of

an ES.A commonly used ES for describing the

magnitude of the difference between twomeans is the difference between meansdivided by the pooled standard deviation, aratio that is referred to as Cohen’s d ( C o h e n ,1988). This ES is calculated by:

(1)

where g = Group.

Although Equation 1 is often used in thecalculation of ES, any of several differentalgebraically equivalent equations can alsobe used (e.g., Cooper and Hedges, 1994). Forthe meta-analysis in this review, ES forindividual studies was calculated using themathematical approach recommended byHunter and Schmidt (1990). Robey andDalebout (1998) described this approach forapplication to hearing, speech, and languageresearch and noted that certain quantitiesmust be able to be extracted clearly from theoriginal studies in order to calculate ES. Inaddition to raw scores or gain scores (meansas well as standard deviations or standarderrors), pr-post test correlation coefficients,and reliability estimates for the outcomemeasures are often needed to perform meta-analyses. When these latter values were notavailable from a study, they were derivedfrom the mean value of those studies thatreported such statistics. When studiesreported outcomes for any one instrument atmore than one measurement interval (e.g., 3,6, or 12 mo. post-hearing-aid intervention),a mean ES was computed for the study as awhole. This was done to minimize a threat tothe validity of a meta-analysis, which canarise from the inclusion of ESs that are notindependent from each other (Robey andDalebout, 1998). However, when differentoutcome measures were used within the sames t u d y, they were treated as independentmeasures. In addition, the meta-analysiswas conducted using a random-effectsapproach because of the heterogeneous natureof the primary studies. Details on differencesbetween fixed-effects and random-effectsmodels of meta-analysis are beyond the scopeof this report, but further explanation isprovided in Cooper and Hedges (1994).

Journal of the American Academy of A u d i o l o g y/ Volume 18, Number 2, 2007

162

Page 13: ASystematic Review of Health-Related Quality of Life and ... · J Am Acad Audiol 18:151–184 (2007) 151 *Department of Communication Sciences and Disorders, University of South Florida,

Quality of Life and Hearing Aids/Chisolm et al

163

The individual ES estimates calculatedfor each of the outcome measures used inthe studies are shown in Table 4. The ESestimates are organized hierarchically basedon whether the data were obtained fromgeneric or disease-specific outcome measures,then by whether the study design providedbetween-subjects data (i.e., experimental vs.control group data from Level 1 RCT or Level2 quasi-experimental designs) or within-subjects data (i.e., pretest vs. posttest datafrom Level 3 nonexperimental designs), andthen by study. For each study, two estimates

of ES are shown: (1) the unadjusted ES (d0)and (2) the adjusted ES for the reliability ofthe outcome measure (d).

An important step in conducting a meta-analysis is to examine the ES estimates thatare obtained for the potential of publicationbias. It is possible that when studies fail toprovide data reaching statistical significance,they do not get published. If this occurs, thenthe studies that are available in the literaturefor research synthesis might yield ESestimates that are positively biased, becausethe studies having negative findings are not

Table 4. Unadjusted Effect Sizes (do) and Effect Sizes Adjusted for the Test-Retest Reliabilities of theOutcome Measure (d) as a Function of HRQoL Domain, Design, (and Levels of Evidence in

Parentheses), and Outcome Measure for the 15 Studies Used in the Meta -analysis

Domain Design Outcome Measure Study do d

Generic Between SELF Mulrow et al (1990) 0.11 0.12(1–2) GDS 0.25 0.27

SPMSQ 0.42 0.47

Within EQ-5D: Anxiety/depression Joore et al (2002) -0.28 -0.33(3) EQ-5D: Mobility -0.07 -0.07

EQ-5D: Self-Care -0.05 -0.06EQ-5D: Pain and Complaints -0.02 -0.03EQ-5D: Daily Activities 0.05 0.06EQ-5D: Visual Analog Scale 0.09 0.10EQ-5D: Daily Activities Joore et al (2003) -0.26 -0.31EQ-5D: Self-Care 0.02 0.03EQ-5D: Pain and Complaints 0.04 0.04EQ-5D: Visual Analog Scale 0.06 0.06EQ-5D: Mobility 0.06 0.07EQ-5D: Anxiety/depression 0.18 0.21MOS SF-36-Social Function 0.23 0.28SPMSQ Mulrow et al (1992a) 0.10 0.11GDS 0.15 0.16MOS SF-36-General Health Stark and Hickson (2004) -0.13 -0.15MOS SF-36-Vitality -0.11 -0.12MOS SF-36-Physical Function -0.06 -0.06MOS SF-36-Bodily Pain -0.05 -0.05MOS SF-36-Mental Health -0.02 -0.02MOS SF-36-Role-Function/Physical 0.03 0.03MOS SF-36-Role-Function Emotional 0.09 0.10MOS SF-36-Social 0.09 0.11

Disease-Specific Between HHIE Abrams et al (1992) 0.93 0.95

(1–2) QDS Mulrow et al (1990) 1.90 2.22Within HHIE Chmiel and Jerger (1996) 0.45 0.46(3) HHIE Dillon et al (1997) 1.35 1.38

HHIE Humes et al (2001) 0.76 0.77HHIE Jerger et al (1996) 0.25 0.26ADP-VAS Joore et al (2002) 1.38 1.52ADP-VAS Joore et al (2003) 1.35 1.49HHIE Malinoff and Weinstein (1989a) 1.68 1.72QDS Mulrow et al (1992a) 0.44 0.52HHIE Mulrow et al (1992a) 0.85 0.87HHIE Newman and Weinstein (1988) 1.65 1.68HHIE Newman et al (1993) 1.06 1.09HHIE Primeau (1997) 1.29 1.32HHIE Stark and Hickson (2004) 1.17 1.20HHIE Taylor (1993) 1.06 1.08

Note: Yueh et al (2001) did not meet the criteria for inclusion.

Page 14: ASystematic Review of Health-Related Quality of Life and ... · J Am Acad Audiol 18:151–184 (2007) 151 *Department of Communication Sciences and Disorders, University of South Florida,

available for inclusion in the analysis. Thus,funnel plots are constructed to check for thepossibility of negative bias in any group ofstudies (Greenhouse and Iyengar, 1994). Ina funnel plot, the sample size is plottedagainst the ES. The funnel or triangularshape formed by the data points indicatesthat, as expected from sampling theory, themagnitude of the ESs obtained from studieshaving small numbers of participants varymuch more than those obtained from studieshaving larger numbers of participants, butthat the ESs for all sample sizes are centeringon a single point on the abscissa (Robey andDalebout, 1998). The funnel plot is examinedto determine if the left angle is missing orsparsely populated. If this occurs, then biasis indicated. Examination of funnel plots forstudies providing ES estimates in the presentmeta-analyses revealed that publication biasdid not influence the results.

The adjusted ES estimates shown inTable 4 are plotted in Figure 2 along withtheir 95% confidence intervals (CIs). Theeffect sizes for each type of data, between- andwithin-subjects, were ordered by theirmagnitude for clarity of presentation inFigure 2. Examination of Table 4 and Figure2 reveals that ES estimates variedconsiderably among the studies with thegreatest differences being for type of HRQoLoutcome measure used. Larger ES estimateswere found for disease-specific than for

generic outcome measures. For each type ofHRQoL measure, ES estimates were largerfor between-subject than for within-subjectanalyses. For many of the generic outcomemeasures, particularly those obtained fromwithin-subject analyses, the CIs for theindividual ES estimates encompassed zero,suggesting that there was no statisticallysignificant effect of hearing aid interventionon HRQoL (i.e., zero is a plausible value for thepopulation ES). In contrast, only two of the ESestimates obtained for disease-specific outcomemeasures yielded CIs encompassing zero.

One of the advantages of conducting ameta-analysis is that the individual ESs,which are weighted for sample size, can beaveraged, thus increasing the accuracy andefficiency with which the populationparameter is estimated. That is, effectsobtained from small samples do not undulyinfluence the calculated value of the average(Robey and Dalebout, 1998). Mean ESs andCIs, as a function of HRQoL and data types,were calculated here using the approachrecommended by Hedges and Olkin (1985).The results are shown in Figure 3. As wouldbe expected from the individual ES estimates,larger mean ESs were obtained when HRQoLwas assessed using disease-specific ratherthan generic outcome instruments.

The benchmarks proposed by Cohen(1988) were used to interpret ES estimatesfor studies using between-subjects analyses

Journal of the American Academy of A u d i o l o g y/ Volume 18, Number 2, 2007

164

Figure 2. Estimated effect size and confidence intervals (vertical bars) for individual studies by design, levelof evidence, and type of outcome measure.

Page 15: ASystematic Review of Health-Related Quality of Life and ... · J Am Acad Audiol 18:151–184 (2007) 151 *Department of Communication Sciences and Disorders, University of South Florida,

Quality of Life and Hearing Aids/Chisolm et al

165

for both the generic and the disease-specificoutcome measures. These benchmarks ared = 0.20, 0.50, and 0.80 for small, medium,and large effects, respectively. Benchmarksfor effects from within-subjects analyses werecalculated using a transformation suggestedby Barcikowski and Robey (1985):

(2)

The average correlation coefficientcalculated for the pretest/posttest datareported in the studies was r = 0.83, resultingin within-subjects benchmarks of d = 0.48,1.21, and 1.95 for small, medium, and largeeffects, respectively. It is important to notethat both the benchmarks recommended byCohen (1988) and the ones calculated forexamining ESs from within-subjects analysisprovide only a starting point for examinationof the effects of hearing aid use on HRQoL.In stating benchmarks, Cohen (1988) madeit clear that the values proposed were selectedto reflect the typical ESs encountered in thebehavioral sciences, and thus they might bemisleading when applied to any one specificarea. In the absence of discipline-specificbenchmarks, however, comparison of a

specific effect, such as that of hearing aidintervention on HRQoL, against the broadcriterion proposed by Cohen (1988), isconsidered a reasonable starting point.

The mean within-subjects ES calculatedfor the generic HRQoL measures was equalto 0.02, with 95% CIs of -0.04 and 0.07. Thisresult suggests that the effect of hearing aiduse on HRQoL is essentially negligible. Recallthat the within-subjects data were obtainedfrom Level 3, nonexperimental studies, andthus should be interpreted with caution.Indeed, the interpretation is different whenthe mean between-subjects ES estimate,which was obtained from higher-level studies(i.e., Level 1 and Level 2), is considered. Themean ES of 0.28 with correspondingconfidence intervals of 0.09 and 0.48 suggeststhat hearing aid use may result in a smalleffect on HRQoL when measured usingdisease-specific instruments.

In contrast to the results obtained withgeneric measures, the mean within-subjectsES estimate of 1.01 with 95% CIs of 0.76 and1.26 obtained for disease-specific outcomes inLevel 3 studies supports a conclusion thathearing aids provide a small to medium effecton HRQoL. As with the mean ES estimateobtained for generic HRQoL measures, whenthe between-subjects data from Level 1 andLevel 2 studies are considered, the mean ESestimate is higher (i.e., 2.07 with 95% CIs of

Figure 3. Estimated mean effect sizes and confidence intervals by research design, level of evidence, and typeof outcome measure.

Page 16: ASystematic Review of Health-Related Quality of Life and ... · J Am Acad Audiol 18:151–184 (2007) 151 *Department of Communication Sciences and Disorders, University of South Florida,

0.51 and 3.63). This result suggests thathearing aids have a robust, medium to largeeffect on HRQoL when outcomes aremeasured using disease-specific instruments.

DISCUSSION

The present investigation was designedto examine the HRQoL benefits of hearing

aids for adults with SNHL by conductingboth qualitative and a quantitative systematicreview of the available research literature. Inthe current era of evidence-based practice,quantitative, meta-analytic, systematicreviews provide a high level of evidence uponwhich health-care decisions can be made.When the question is one about treatmente f f i c a c y, systematic reviews would ideallylimit studies for inclusion to those using RCTdesigns. Of all the experimental designs,R C Ts have the least likelihood of introducingany biases that can influence results,particularly when double blinding is used. A sdemonstrated in the present review, hearingaid research has typically not employed RCTs .In fact, only two of the studies included hereused an RCT design, and blinding was notemployed in any of the studies. Generally, thehearing aid literature reveals that quasi-experimental designs are commonly used.However, although several studies includedin this review employed quasi-experimentaldesigns, only one used that methodology toaddress the question of concern here. In fact,most of the included studies were judged tobe nonexperimental in nature, because theylacked appropriate control groups, whichincreased the likelihood for bias and limitedthe usefulness of their findings. A l t h o u g hthe data from these latter studies are stilluseful, the potential for bias can limit makingstrong recommendations from them,especially if they were the only study designsfound. Therefore, in order to strengthen therecommendations that can be made regardingthe benefits of HRQoL from the use of hearingaids, future research in this area shouldstrive to include appropriate control groupsand the RCT as the optimal strategy.

In addition to the level of evidence andexperimental designs employed, qualityassessment of studies is important whenconsidering their contribution to the evidencefor making clinical decisions about aparticular treatment. Quality assessmentrevealed that the studies included in this

systematic review had both weaknesses andstrengths. For example, weaknesses in mostof the studies involved their failure to includecontrol groups, power analyses to determineappropriate sample sizes, and adequatediscussions of dropouts. Strengths observedin the studies involved descriptions of theirinclusion/exclusion criteria for participants,hearing aid fitting procedures, verificationsof hearing aid fittings, use of appropriatestatistical analyses, and, when control groupswere used, equivalence of the groups atbaseline. Again, attending to thesemethodological considerations wouldstrengthen future studies.

Although some caution should be takenin interpreting the overall results of thissystematic review and meta-analysis, thefindings can be helpful in makingrecommendations about treatments for adultswith SNHL. The disease-specific HRQoLmeasures complemented the qualitativefindings of this systematic review anddemonstrated a clear and robust positivebenefit of hearing aid use. The ES estimatesand CIs were in the medium-to-large rangefor studies using control groups, and themean ES estimate was in the medium rangefor those lacking control groups. Thus, thepositive findings from both the qualitative andquantitative synthesis of the studiesmeasuring disease-specific HRQoL o u t c o m e sare important, because they show thathearing aid use improves the psychological,social, and emotional well-being of adultswith acquired SNHL.

The outcomes for hearing aidintervention as measured by genericinstruments were also important becausethese tools can be used to compare treatmentsfor different diseases and disorders. However,because of their broad applicability, outcomesmeasured with generic instruments are oftennot as robust as those observed from disease-specific instruments. Indeed, in the body ofliterature reviewed here, fewer studiesreported statistically significant changes inHRQoL as a function of hearing aid use forgeneric than for disease-specific measures.Further, examination of the correspondingESs for within-subjects effects on genericmeasures generally revealed that they failedto provide evidence to support a strongconclusion that hearing aids improve HRQoL.Even so, it was encouraging that the fewstudies that provided potentially less biased

Journal of the American Academy of A u d i o l o g y/ Volume 18, Number 2, 2007

166

Page 17: ASystematic Review of Health-Related Quality of Life and ... · J Am Acad Audiol 18:151–184 (2007) 151 *Department of Communication Sciences and Disorders, University of South Florida,

Quality of Life and Hearing Aids/Chisolm et al

167

results due to their inclusion of control groups(i.e., between-subjects effects) had mean ESestimates for generic HRQoL that were atleast in the small range.

Considering the known negative impactsthat hearing loss has on adults’ quality of life(Bess et al, 1989; Mulrow et al, 1990; Kelleret al, 1999; National Council on the Aging,1999; Strawbridge et al, 2000; Dalton et al,2003; Pugh, 2004), it is reasonable to questionwhy so few studies provide evidence ofpositive treatment effects resulting from theuse of hearing aids when measured by genericinstruments. In discussing this issue, A b r a m sand colleagues (2005) pointed out that mostof the available generic health statusinstruments fail to include any questionsthat are directly related to hearing and oralcommunication. Earlier, Bess (2000) noted thelack of sensitivity to hearing aid interventionby many of the generic HRQoL instrumentsused in the studies included in this systematicreview.

Future studies should consider usingnew outcome measures that are both sensitiveto the consequences of hearing loss on HRQoLand responsive to interventions for it. Forexample, McArdle and colleagues (2005)examined the use of the World HealthOrganization-Disability Assessment Schedule(WHO-DAS II; World Health Organization,1999), which is a multidimensional toolcontaining two domain scores (communicationand participation) that are specifically relatedto the consequences of hearing loss andmeasuring the outcomes of hearing aidintervention. McArdle et al (2005) examinedthe effects of hearing aid intervention on theWHO-DAS II composite or total score, usinga randomized-controlled research design ina study having 380 participants. Theyreported a statistically significant treatmenteffect and an ES estimate of d = 0.20, whichis further evidence that hearing aid use doesprovide a measurable and consistent, albeitsmall, effect in terms of generic HRQoL.

Although this quantitative systematicreview revealed that hearing aids provideonly a small effect on generic HRQoL, itshould not be concluded that this might be thecase for all patients with hearing loss acrossall rehabilitative contexts. Recall that thestudies included in this review involvedparticipants with different ages, degrees ofhearing loss, hearing aid experience, andhealth-care delivery systems. Moreover,

considering the National Council on the A g i n g(1999) report of untreated hearing loss beingassociated with depression, anxiety, and socialisolation, hearing aids may have a greaterimpact on generic HRQoL for some patientsand their families than for others (Van Vliet,2005). For example, hearing aids may reducedepression in some elderly persons or in thosewith severe degrees of communicationimpairment. Future studies should focus onassessing the effects of hearing aid treatmenton generic HRQoL using homogeneous groupsof participants comparing new versusexperienced hearing aid users with specificpersonality profiles in similar rehabilitativecontexts and carefully selected outcomemeasures (Cox, 2005). In addition, futureresearch should build on the findings fromthis investigation for deriving contextuallyappropriate benchmarks to use in audiologictreatment (Cohen, 1988; Kline, 2004).

The findings of this quantitativesystematic review are important for at leastthree reasons. First, conducting systematicreviews with meta-analyses and computingESs improves upon traditional approaches tointervention research that use null-hypothesis significance testing to assesswhether a particular result is statisticallysignificant (Cox, 2005). Null-hypothesissignificance testing assesses whether aparticular result is statistically significant,with significance translating to possibleevidence supporting the use of a given typeof intervention. However, statisticalsignificance is dependent on factors likesample size and does not always translate intoclinical significance for use in EBP. The useof ESs emphasizes the magnitude of thefinding independent of sample size (Cox,2005). Therefore, audiologists can nowconfidently say that the use of hearing aidsgreatly enhances patients’ d i s e a s e - s p e c i f i cH R Q o L by reducing the psychological, social,and emotional effects of SNHL, and, further,that the strongest evidence supports aconclusion that hearing aids have a positiveinfluence on generic aspects of HRQoL.

Second, in an era of increasedaccountability and shrinking health-careresources, the results of this quantitativesystematic review provide evidence foradministrators, third-party payers, andadults with hearing impairment and theirfamilies that hearing aid use impactspositively on HRQoL. Third-party payers

Page 18: ASystematic Review of Health-Related Quality of Life and ... · J Am Acad Audiol 18:151–184 (2007) 151 *Department of Communication Sciences and Disorders, University of South Florida,

often require evidence about the efficacy of medical interventions prior toreimbursement. Although results fromdisease-specific HRQoL outcome measurespreclude comparisons of hearing aids to otherinterventions, the findings of the meta-analysis serve as convincing evidence of thequality of life enhancement derived from theuse of amplification.

Third, the results of this quantitativesystematic review revealed that ourprofession has a relatively good arsenal ofdisease-specific HRQoL outcome measuresthat are highly responsive to interventionsthat are designed to manage SNHL amongheterogeneous patient populations (Deyo andPatrick, 1989). A l t e r n a t i v e l y, there appear tobe few generic HRQoL outcome measuresavailable that are likely to be responsive toaspects of hearing aid intervention.

Some possible limitations of thesystematic review process should beaddressed before drawing conclusions fromthis report. These limitations includepublication bias, heterogeneity of the studiesincluded in the meta-analysis, the time-sensitive nature of the search and retrievalprocess, and weaknesses of the individualstudies. As discussed previously, publicationbias can result from the tendency for journalsto publish only statistically significantfindings. Recall that the search and retrievalprocess used in this review did include handsearching for unpublished studies inproceedings of recent conferences and studiesreported in non-peer-reviewed journals.H o w e v e r, a decision was made to focus onstudies published in peer-reviewed journals toenhance the chances of finding support at thehighest levels of evidence. Although theexclusion of unpublished data could havebiased the results of this systematic review,examination of the funnel plot for the extractedES measures suggested that it did not.

A good systematic review assesses for thehomogeneity of the findings of individualstudies that are included in a meta-analysis.I d e a l l y, meta-analyses are conducted withstudies having similar characteristics (i.e.,designs, participant samples, application ofinterventions, etc.) and that do not differ inclinically significant ways. Traditional fixed-effects meta-analyses test for heterogeneityusing the Q statistic, which attempts toisolate studies that have homogeneousfindings and exclude outlier studies. As stated

e a r l i e r, the meta-analysis completed herewas conducted using a random-effects ratherthan a fixed-effects approach because of theheterogeneous nature of the studies. Therandom-effects approach accounts for thedifferences among all studies whendetermining the magnitude of an effect.Therefore, the heterogeneity of the studieswas accounted for through use of thisstatistical procedure. Future studies shouldfocus on more homogeneous patientpopulations for assessing generic HRQoL.

Another limitation of systematic reviewsof the published literature is the time-sensitive nature of the search and retrievalprocess. For example, the search and retrievalprocess for this systematic review wascompleted in August 2004. As noted earlier,at least one additional study (i.e., McArdle etal, 2005) that would have met the inclusioncriteria has been published since then.U n f o r t u n a t e l y, the time cutoff for thissystematic review precluded evidencepublished after August 2004 from beingincluded here. Thus, systematic reviewsshould be updated regularly in order to insureinclusion of the most up-to-date informationon a topic. Despite their ability to providehighly relevant and useful aggregatedinformation on a given topic, publishedsystematic reviews are static and can onlypresent the state of evidence for interventionsup to a particular point in time.

Finally, as discussed earlier, systematicreviews only reflect the inherent strengthsand weaknesses of the included studies. Onlytwo of the studies included here used RCTs,while most of the investigations used pre-posttest designs. Many of the studies that werereviewed here failed to (1) conduct poweranalyses specifying adequate sample sizes, (2)randomize participants to treatment andcontrol groups, and (3) discuss drop-outs.Investigators addressing the efficacy ofaudiologic treatment should seek to designhigh quality, randomized, double-blinded,controlled clinical trials according torecommended guidelines (CONSORT, 2006a,2006b). In addition, researchers couldfacilitate future systematic reviews if theywould consistently and fastidiously developarticle titles that accurately reflect the truenature of their studies, provide well-constructed abstracts that reveal criticalaspects about the design of the study, usewell-selected key words that lead to easy

Journal of the American Academy of A u d i o l o g y/ Volume 18, Number 2, 2007

168

Page 19: ASystematic Review of Health-Related Quality of Life and ... · J Am Acad Audiol 18:151–184 (2007) 151 *Department of Communication Sciences and Disorders, University of South Florida,

Quality of Life and Hearing Aids/Chisolm et al

169

retrieval and evaluation, and report ESs andCIs for their results (CONSORT, 2006a,2006b). Only studies that provide results atthe highest levels of evidence can permitinferences about causality between audiologictreatment and improved outcomes for adultswith SNHL.

One way to summarize the results of anysystematic review of the literature, whilekeeping the cautions discussed above in mind,is to assign a grade to the health-carerecommendations that result from theevidence. Grading provides an indicator of theextent to which the evidence supports aparticular recommendation. Cox (2005)discussed how recommendations fromsystematic reviews can be used for clinicaldecision making by assigning them grades of“A,” “B,” “C,” or “D.” For example, a grade of“A” suggests that clinicians can be veryconfident in making the recommendation fora particular intervention, because it issupported by a body of high-quality relevantresearch. However, a grade of “D” means therecommendation must be made with greatcaution.

The findings of this quantitativesystematic review suggest that arecommendation grade of “B” appears to bewarranted for the use of hearing aids toimprove adults’ H R Q o L considering the levelsof evidence and quality of the includedstudies. Moreover, the average ES estimatesand CIs discerned from the meta-analysissupport a recommendation for the use ofhearing aids to improve adults’ H R Q o L .Improvement is the most likely outcome,particularly when hearing related effects aredirectly assessed.

As stated earlier, EBP refers to anapproach by which current, high-qualityresearch evidence is integrated withpractitioner expertise and patient preferencesand values into the process of making clinicaldecisions (American Speech-Language-Hearing Association, 2004). Systematicreviews with meta-analyses are becomingmore common in the audiologic literatureand can assist patients, audiologists, primary-care physicians, and other health-careprofessionals in selecting the mostappropriate treatment options for SNHL.H o w e v e r, unlike other chronic healthconditions with multiple treatmentalternatives, the only viable option for mostcases of SNHL is the use of hearing aids,

which relegates clinical decision making toone of whether to pursue treatment. In doingso, patients and their health-careprofessionals must weigh the risks andbenefits of pursuing amplification, acomparatively noninvasive, low-risktreatment with considerable potentialbenefits. Indeed, most states now requiretrial periods for hearing aids so that patientsface little financial risk if they are notcompletely satisfied with the results of theirpurchase. Therefore, the modest evidence ofbenefits in HRQoL provided by thissystematic review become quite powerfulwhen considering that hearing aid use is theonly viable treatment for SNHL, a conditionwith insidious, potentially devastating effectswhen left untreated (National Council on theAging, 1999).

CONCLUSIONS

This systematic review with meta-analysisresulted in the following conclusions:• Hearing aid use (a comparatively

noninvasive, low-risk option withconsiderable potential benefits,which is the only viable treatmentfor SNHL) improves adults’ H R Q o Lby reducing psychological, social,and emotional effects of SNHL, aninsidious, potentially devastatingchronic health condition if leftunmanaged.

• The quantitative systematic reviewprocess provided a powerful methodfor assessing the HRQoL benefits ofamplification; however, itsconclusions are only as robust asthe studies that are included in ther e v i e w, and it is a time-sensitiveendeavor that needs to be updatedperiodically in order to reveal thebest and most current evidence forparticular treatments.

• Although the field of audiologyappears to have a sufficient batteryof disease-specific tools, it shouldstrive to use, adapt, or developgeneric instruments that aresensitive to and appropriate forassessing changes in hearing aidusers’ and their families’ HRQoL asa result of amplification.

• Researchers should exercise great

Page 20: ASystematic Review of Health-Related Quality of Life and ... · J Am Acad Audiol 18:151–184 (2007) 151 *Department of Communication Sciences and Disorders, University of South Florida,

care in designing, conducting, andreporting their studies in order tomaximize their contributions to EBP.

• Future research in this area shouldstrive to use RCT designs andgeneric HRQoL measures that aresensitive to the effects of andtreatments for hearing loss.Investigators should conduct poweranalyses, employ both experimentaland control groups, use doubleblinding, adequately describeparticipant inclusion/exclusioncriteria, provide intention-to-treatanalyses, discuss dropouts, andcompute ESs and CIs for statisticallysignificant results wheneverpossible.

• The audiologic community, patientswith hearing loss and their families,physicians and other health-careproviders, and third-party entitiesshould be encouraged that hearingaids can provide considerableHRQoL benefits for the increasingnumbers of the population havingSNHL.

A c k n o w l e d g m e n t s . Funding for this work was pro-vided in part by the American Academy of A u d i o l o g y.The members of the Health-Related Quality of LifeBenefits of Amplification in Adults Task Force wishto thank Angela Loavenbruck, Past President of theAmerican Academy of A u d i o l o g y, for her vision inestablishing this group and our mission. We wouldalso like to thank Jeff Kromrey, Ph.D., for his assis-tance in the completion of the meta-analysis.A d d i t i o n a l l y, we would like to thank Lindsey Latiolaisand Amber Reith for their assistance with compilingthe references for this review.

REFERENCES

Abrams H, Chisolm TH, McArdle R. (2005) Health-related quality of life and hearing aids: a tutorial.Trends Amplif 9:99–109.

Abrams H, Hnath-Chisolm T. (2000) Outcome meas-ures: the audiological difference. In: Hosford-DunnH, Roeser RJ, Valente M, eds. Audiology: PracticeManagement. New York: Thieme, 69–95.

*Abrams HB, Hnath-Chisolm T, Guerreiro SM,Ritterman SI. (1992) The effects of intervention strat-egy on self-perception of hearing handicap. Ear Hear13:371–377.

Agency for Healthcare Research and Quality. (2002)Rating the Strength of Scientific Research Findings.AHRQ Publication No. 02-P022. www. a h r q . g o v /clinic/epcsums/strenfact.htm (accessed February 12,2005).

American Speech-Language-Hearing A s s o c i a t i o n .(2004) Evidence-Based Practice in CommunicationDisorders: An Introduction [ Technical Report].h t t p : / / w w w. a s h a . o r g / m e m b e r s / d e s k r e f -journals/deskref/default (accessed December 12, 2005).

Andersson G, Green M. (1995) Anxiety in elderly hear-ing impaired persons. Percept Mot Skills 8 1 : 5 5 2 – 5 5 4 .

Barcikowski RS, Robey RR. (1985) Sample size selec-tion in the single group repeated measures analysis.Paper presented at the Annual Convention of theAmerican Educational Research Association, Chicago.

Bess FH. (2000) The role of generic health-relatedquality of life measures in establishing audiologicalrehabilitation outcomes. Ear Hear 21:74S–79S.

Bess FH, Lichtenstein MJ, Logan SA. (1990) Makinghearing impairment functionally relevant: linkageswith hearing disability and handicap. Acta OtolaryngolSuppl 476:226–231.

Bess FH, Lichtenstein MJ, Logan SA, Burger MC,Nelson E. (1989) Hearing impairment as a determi-nant of function in the elderly. J Am Geriatr Soc37:123–128.

Bruley DK. (1999) Beyond reliability and validity:analysis of selected quality-of-life instruments for usein palliative care. J Palliat Med 2:299–309.

Campbell VA, Crews JE, Moriarty DG, Zack MM,Blackman DK. (1999) Surveillance for sensory impair-ment, activity limitation, and health-related qualityof life among older adults – United States, 1993–1997.MMWR CDC Surveill Summ 48:131–156.

*Chmiel R, Jerger J. (1996) Hearing aid use, centralauditory disorder, and hearing handicap in elderlypersons. J Am Acad Audiol 7:190–202.

Cohen J. (1988) Statistical Power Analysis for theBehavioral Sciences, 2nd Edition. Hillsdale, NJ:Erlbaum Associates.

C O N S O R T- Strength in Science, Sound Ethics. (2006a)Consort E-checklist. w w w. c o n s o r t - s t a t e m e n t . o r g /Downloads/download.html (accessed June 30, 2006).

C O N S O R T- Strength in Science, Sound Ethics. (2006b)Consort E-flowchart. w w w. c o n s o r t - s t a t e m e n t . o r g /Downloads/download.html (accessed June 30, 2006).

Cooper H, Hedges LV. (1994) The Handbook ofResearch Synthesis. New York: Russell SageFoundation.

Cox RM. (2005) Evidence-based practice in provisionof amplification. J Am Acad Audiol 19:419–438.

Dalton DS, Cruickshanks KJ, Klein BE, Klein R,Wiley TL, Nonhdahl DM. (2003) The impact of hear-ing loss on quality of life in older adults. G e r o n t o l o g i s t43:661–668.

Deyo RA, Patrick DL. (1989) Barriers to the use ofhealth status measures in clinical investigation,patient care, and policy research. Med Care27:S254–S268.

*Dillon H, James A, Ginis J. (1997) Client OrientedScale of Improvement (COSI) and its relationship to

Journal of the American Academy of A u d i o l o g y/ Volume 18, Number 2, 2007

170

Page 21: ASystematic Review of Health-Related Quality of Life and ... · J Am Acad Audiol 18:151–184 (2007) 151 *Department of Communication Sciences and Disorders, University of South Florida,

Quality of Life and Hearing Aids/Chisolm et al

171

several other measures of benefit and satisfactionprovided by hearing aids. J Am Acad Audiol 8 : 2 7 – 4 3 .

E u r o Q o L Group. (1990) Euroqol- a new facility forthe measurement of health related quality of life.Health Policy 16:199–208.

Fifer RC, Jerger JF, Berlin CI, Tobey EA, CampbellJC. (1983) Development of a dichotic sentence iden-tification test for hearing-impaired adults. Ear Hear4:300–305.

Greenhouse JB, Iyengar S. (1994) Sensitivity analy-sis and diagnostics. In: Cooper H, Hedges LV, eds.The Handbook of Research Synthesis. New Yo r k :Russell Sage Foundation, 383–398.

Healthy People 2010. (2004) Hearing Health ProgressR e v i e w. h t t p : / / w w w. n i d c d . n i h . g o v / h e a l t h / h e a l t h y-h e a r i n g / w h a t _ h h / p r o g r e s s _ r e v i e w _ 0 4 . a s p # r e f 2 3 .(accessed March 19, 2005).

Hedges LV, Olkin I. (1985) Statistical Methods forM e t a - a n a l y s i s . Burlington, MA: Elsevier Science andTechnology Books.

Higgins JPT, Green S. (2005) Cochrane Handbook forSystematic Reviews of Interventions 4.2.5. [Updatedmay 2005] In: Cochrane Library3. Chichester, UK:Wiley.

*Humes LE, Garner CB, Wilson DL, Barlow NN.(2001) Hearing-aid outcome measured following onemonth of hearing aid use by the elderly. J SpeechLang Hear Res 44:469–486.

Hunter JE, Schmidt FL. (1990) Methods of Meta-analysis: Correcting Error and Bias in ResearchFindings. London: Sage.

*Jerger J, Chmiel R, Florin E, Pirozzolo F, Wilson N.(1996) Comparison of conventional amplification andan assistive listening device in elderly persons. EarHear 17:490–504.

*Joore MA, Brunenberg DE, Chenault MN, A n t e u n i sLJ. (2003) Societal effects of hearing aid fitting amongthe moderately hearing impaired. Int J Audiol42:152–160.

*Joore MA, Potjewijd J, Timmerman AA, A u t e u n i sLJ. (2002) Response shift in the measurement of qual-ity of life in hearing impaired adults after hearingaid fitting. Qual Life Res 11:299–307.

Keller BK, Morton JL, Thomas VS, Potter JF. (1999)The effect of visual and hearing impairments on func-tional status. J Am Geriatr Soc 47:1319–1325.

Kiessling J, Pichora-Fuller MK, Gatehouse S,Stephens D, Arlinger S, Chisolm T, Davis AC, ErberN P, Hickson L, Holmes A, Rosenhall U, von Wedel H.(2003) Candidature for and delivery of audiologicalservices: special needs of older people. Int J Audiol42(Suppl.):2S92–101.

Kline RB. (2004). Beyond Significance Te s t i n g :Reforming Data Analysis Methods in BehavioralSciences. Washington, DC: American PsychologicalAssociation.

Kramer SE, Kapteyn TS, Kuik DJ, Deeg DJ. (2002)The association of hearing impairment and chronic

diseases with psychosocial health status in older age.J Aging Health 14:122–137.

Linn MW, Linn BS. (1984) Self-evaluation of life func-tion (self) scale: a short, comprehensive self-report ofhealth for elderly adults. J Gerontol 39:603–612.

Malinoff RL, Weinstein BE. (1989a) Measurement ofhearing aid benefit in the elderly. Ear Hear10:354–356.

McArdle R, Chisolm TH, Abrams HB, Wilson RH,Doyle PJ. (2005) The WHO-DAS II: measuring out-comes of hearing aid intervention for adults. TrendsAmplif 9:127–143.

McCallum J. (1995) The SF-36 in an A u s t r a l i a nsample: validating a new, generic health status meas-ure. Aust J Public Health 19:160–166.

*Mulrow CD, Aguilar C, Endicott JE, Tuley MR, Ve l e zR, Charlip WS, Rhodes MC, Hill JA, DeNino LA.(1990) Quality-of-life changes and hearing impair-ment: a randomized trial. Ann Intern Med113:188–194.

Mulrow CD, Tuley MR, Aguilar C. (1992a) Sustainedbenefits of hearing aids. J Speech Lang Hear Res35:1402–1405.

National Council on the Aging. (1999) T h eConsequences of Untreated Hearing Loss in OlderPersons [Study conducted by the Seniors ResearchGroup, an alliance between The National Council onthe Aging and Market Strategies Inc.]. Washington,DC: National Council on the Aging.

National Institutes of Health. (1993) Quality of LifeAssessment: Practice, Problems, and Promise -Proceedings of a Workshop, October 15–17, 1990.Washington, DC: U.S. Department of Health andHuman Services.

*Newman CW, Hug GA, Wharton JA, Jacobson GP.(1993) The influence of hearing aid cost on perceivedbenefit in older adults. Ear Hear 14:285–289.

*Newman CW, Weinstein BE. (1988) The HearingHandicap Inventory for the Elderly as a measure ofhearing aid benefit. Ear Hear 9:81–85.

Newman CW, Weinstein BE, Jacobson GP, Hug G.(1990) The Hearing Handicap Inventory for Adults:psychometric adequacy and audiometric correlates.Ear Hear 11:430–433.

Pfeiffer E. (1975) A short portable mental status ques-tionnaire for the assessment of organic brain deficitin elderly patients. J Am Geriatr Soc 23:433–441.

*Primeau RL. (1997) Hearing aid benefit in adultsand older adults. Semin Hear 18:29–36.

Pugh KC. (2004) Health status attributes of olderAfrican-American adults with hearing loss. J NatlMed Assoc 96:772–779.

Robey R. (2004) Levels of evidence. The ASHA L e a d e r(April 13):5.

Robey RR, Dalebout SD. (1998) A tutorial on con-ducting meta-analyses of clinical outcome research.J Speech Lang Hear Res 41:1227–1241.

Page 22: ASystematic Review of Health-Related Quality of Life and ... · J Am Acad Audiol 18:151–184 (2007) 151 *Department of Communication Sciences and Disorders, University of South Florida,

Robey RR, Schultz MC. (1993) Optimizing Theoriesand Experiments. San Diego, CA: Singular.

Schulz KF, Chalmers I, Hayes RJ, Altman DG. (1995)Empirical evidence of bias. Dimensions of method-ological quality associated with estimates of treatmenteffects in controlled trials. JAMA 273:408–412.

Scottish Intercollegiate Guidelines Network. (2001)SIGN 50: A Guideline Developers’ H a n d b o o k .h t t p : / / w w w. s i g n . a c . u k / g u i d e l i n e s / f u l l t e x t / 5 0 / i n d e x . h t ml (accessed March 8, 2004).

*Stark L, Hickson P. (2004) Outcomes of hearing aidfitting for older people with hearing impairment andtheir significant others. Int J Audiol 43:390–398.

Strawbridge WJ, Wallhagen MI, Shema SJ, KaplanGA. (2000) Negative consequences of hearing impair-ment in old age: a longitudinal analysis. G e r o n t o l o g i s t40:320–326.

* Taylor KS. (1993) Self-perceived and audiometricevaluations of hearing aid benefit in the elderly. EarHear 14:390–394.

Tuley MR, Mulrow CD, Aguilar C, Velez R. (1990) Acritical reevaluation of the Quantified Denver Scaleof Communication Function. Ear Hear 11:56–61.

Uhlmann RF, Larson EB, Rees TS, Koepsell TD,Duckert LG. (1989) Relationship of hearing impair-ment to dementia and cognitive dysfunction in olderadults. JAMA 261:1916–1919.

Van Vliet D. (2005) The current status of hearingcare: can we change the status quo? J Am Acad Audiol16:410–418.

Ventry IM, Weinstein IM. (1982) The hearing hand-icap inventory for the elderly: a new tool. Ear Hear3:128–134.

Ware JE Jr, Sherbourne CD. (1992) The MOS 36-itemshort-form health survey (SF-36): I. Conceptual frame-work and item selection. Med Care 30:473–483.

Ware JE, Snow KK, Kosinski M, Gandek B. (1993)SF-36 Health Survey Manual and InterpretationGuide. Boston: Health Institute.

Wayner DS, Abrahamson JE. (2001) Learning to HearAgain, Second Edition. Latham, NY: Hear A g a i nPublishing.

Weinstein BE, Ventry IM. (1982) Hearing impair-ment and social isolation in the elderly. J Speech HearRes 25:593–599.

World Health Organization. (1999) The World HealthOrganization Disability Assessment Schedule PhaseII Field Trial Instrument. Geneva, Switzerland: Wo r l dHealth Organization.

World Health Organization. (2004) I n t e r n a t i o n a lStatistical Classification of Diseases and HealthRelated Problems ICI-10, Tenth Edition. G e n e v a ,Switzerland: World Health Organization.

Yesavage JA, Brink TL, Rose TL, Lum O, Huang V,Adey M, Leirer VO. (1982–83) Development and val-idation of a geriatric depression screening scale: apreliminary report. J Psychiatr Res 17:37–49.

* Yueh B, Souza PE, McDowell JA, Collins MP, LoovisC F, Hendrick SC, Ramsey SD, Deyo RA. (2001)Randomized trial of amplification strategies. A r c hOtolaryngol Head Neck Surg 127:1197–1204.

*Included studies

APPENDIX A. EXCLUDED STUDIES

Baumfield A, Dillon H. (2001) Factors affecting theuse and perceived benefit of ITE and BTE hearingaids. Br J Audiol 35:247–258.

Beamer SL, Grant KW, Walden BE. (2000) Hearingaid benefit in patients with high-frequency hearingloss. J Am Acad Audiol 11:429–437.

Berninger E, Karlsson KK. (1999) Clinical study ofWidex Senso on first-time hearing aid users. ScandAudiol 28:117–125.

Bridges JA, Bentler RA. (1998) Relating hearing aiduse to well-being among older adults. Hear J51(7):39–44.

Brooks DN, Hallam RS. (1998) Attitudes to hearingdifficulty and hearing aids and the outcome of audi-ological rehabilitation. Br J Audiol 32:217–226.

Brooks DN, Hallam RS, Mellor PA. (2001) The effectson significant others of providing a hearing aid to thehearing-impaired partner. Br J Audiol 35:165–171.

Cacciatore F, Napoli C, Abete P, Marciano E, TriassiM, Rengo F. (1999) Quality of life determinants andhearing function in an elderly population: OsservatorioGeriatrico Compano Study Group. G e r o n t o l o g y45:323–328.

Cord MT, Leek MR, Walden BE. (2000) Speech recog-nition ability in noise and its relationship to perceivedhearing aid benefit. J Am Acad Audiol 11:475–483.

Cox RM, Alexander GC. (1995) The abbreviated pro-file of hearing aid benefit. Ear Hear 16:176–186.

Cox RM, Alexander GC, Beyer CM. (2003) Norms forthe international outcome inventory for hearing aids.J Am Acad Audiol 14:403–413.

Crowley HJ, Nabelek IV. (1996) Estimation of client-assessed hearing aid performance based upon unaidedvariables. J Speech Hear Res 39:19–27.

Cunningham DR, Williams KJ, Goldsmith LJ. (2001)Effects of providing and withholding postfitting fine-tuning adjustments on outcome measures in novicehearing aid users: a pilot study. Am J Audiol 1 0 : 1 3 – 2 3 .

Dillon H, Koritschoner E, Battaglia J, Lovegrove R,Ginis J, Mavrias G, Carnie L, Ray P, Forsythe L,Towers E, Goulios H, Macaskill F. (1991)Rehabilitation effectiveness II: Assessing the out-comes for clients entering a national hearingrehabilitation program. Aust J Audiol 13:68–82.

Dye CJ, Peak MF. (1983) Influence of amplificationon the psychological functioning of older adults withneurosensory hearing loss. J Acad Rehabil Audiol16:210–220.

Journal of the American Academy of A u d i o l o g y/ Volume 18, Number 2, 2007

172

Page 23: ASystematic Review of Health-Related Quality of Life and ... · J Am Acad Audiol 18:151–184 (2007) 151 *Department of Communication Sciences and Disorders, University of South Florida,

Quality of Life and Hearing Aids/Chisolm et al

173

Gatehouse S. (1994) Components and determinantsof hearing aid benefit. Ear Hear 15:30–49.

Gatehouse S. (1999) A self-report outcome measurefor the evaluation of hearing aid fittings and service.Health Bull Edinb 57:424–436.

Gatehouse S, Noble W. (2004) The Speech, Spatialand Qualities of Hearing Scale (SSQ). Int J Audiol43:100–114.

Haggard MP, Foster JR, Iredale FE. (1981) Use andbenefit of postaural aid in sensory hearing loss. S c a n dAudiol 10:45–52.

Harless EL, McConnell F. (1982) Effects of hearingaid use on self concept in older persons. J Speech HearDisord 47:305–309.

Haskell GB, Noffsinger D, Larson VD, Williams DW,Dobie RA, Rogers JL. (2002) Subjective measures ofhearing aid benefit in the NIDCD/VA Clinical Trial.Ear Hear 23:301–307.

Horwitz AR, Turner CW. (1997) The time course ofhearing aid benefit. Ear Hear 18:1–11.

Humes LE. (2001) Issues in evaluating the effec-tiveness of hearing aids in the elderly: what tomeasure and when. Semin Hear 22:303–315.

Humes LE. (2003) Modeling and predicting hearingaid outcome. Trends Amplif 7:41–75.

Humes LE, Christensen LA, Bess FH, Hedley-Williams A. (1997) A comparison of the benefitprovided by well-fit linear hearing aids and instru-ments with automatic reductions of low-frequencygain. J Speech Lang Hear Res 40:666–685.

Humes LE, Wilson DL. (2003) An examination ofchanges in hearing-aid performance and benefit inthe elderly over a 3-year period of hearing-aid use. JSpeech Lang Hear Res 46:137–145.

Humes LE, Wilson DL, Barlow NN, Garner CB, A m o sN. (2002b) Longitudinal changes in hearing aid sat-isfaction and usage in the elderly over a period of oneor two years after hearing aid delivery. Ear Hear23:428–438.

Humes LE, Wilson DL, Barlow NN, Garner C. (2002a)Changes in hearing-aid benefit following 1 or 2 yearsof hearing-aid use by older adults. J Speech LangHear Res 45:772–782.

Humes LE, Wilson DL, Humes AC. (2003)Examination of differences between successful andunsuccessful elderly hearing aid candidates matchedfor age, hearing loss and gender. Int J Audiol42:432–441.

Hutton CL, Canahl JA. (1985) Scaling patient reportsof hearing aid benefit. J Aud Res 25:255–269.

Jerram JC, Purdy SC. (1997) Evaluation of hearingaid benefit using the Shortened Hearing A i dPerformance Inventory. J Am Acad Audiol 8:18–26.

Jerram JC, Purdy SC. (2001) Te c h n o l o g y, expecta-tions, and adjustment to hearing loss: predictors ofhearing aid outcome. J Am Acad Audiol 12:64–79.

Kochkin S. (1997) What is the viable market for hear-ing aids? Hear J 50(1):31–39.

Kochkin S. (1998) Correlates of hearing aid purchaseintent. Hear J 51(1):32–41.

MacKenzie K, Browning GG. (1991) Randomized,cross over study to assess patient preference for anacoustically modified hearing aid system. J LaryngolOtol 105:405–408.

Mäki-Torkko EM, Brorsson B, Davis A, Mair LWS,Myhre KL, Roine RP, Rosenhall U, Sorri MJ,Stilvén S. (2001) Hearing impairment among adults– extent of the problem and scientific evidence on theoutcome of hearing aid rehabilitation. Scand AudiolSuppl 54:8S–15S.

Malinoff RL, Weinstein BE. (1989b) Changes in self-assessment of hearing handicap over the first yearof hearing aid use. J Acad Rehabil Audiol 22:54–60.

Marttila TI, Jauhiainen T. (1996) Hearing disabilityassessment in evaluating hearing aid benefit. S c a n dAudiol 25:121–125.

McCarthy P. (1996) Hearing aid fitting and audio-logic rehabilitation a complementary relationship.Am J Audiol 5:24–28.

Mulrow CD, Tuley MR, Aguilar C. (1992b) Correlatesof successful hearing aid use in older adults. Ear Hear13:108–113.

Newman CW, Weinstein BE. (1989) Test-retest reli-ability of the Hearing Handicap Inventory for theElderly using two administration approaches. E a rHear 10:190–191.

Newman CW, Weinstein BE, Jacobson GP, Hug GA.(1991) Test-retest reliability of the hearing handicapinventory for adults. Ear Hear 12:355–357.

Norman M, George CR, McCarthy D. (1994) The effectof pre-fitting counselling on the outcome of hearingaid fittings. Scand Audiol 23:257–263.

Owens E, Fujikawa S. (1980) The hearing perform-ance inventory and hearing aid use in profoundhearing loss. J Speech Hear Res 23:470–479.

Palmer CV, Adams SW, Bourgeois M, Durrant J, RossiM. (1999) Reduction in caregiver-identified problembehaviors in patients with Alzheimer disease post-hearing-aid fitting. J Speech Lang Hear Res42:312–328.

Parent TC, Chmiel R, Jerger J. (1997) Comparisonof performance with frequency transposition hearingaids and conventional hearing aids. J Am Acad Audiol8:355–365.

Parving A, Parving I, Erlendsson A, Christensen B.(2001) Some experiences with hearing disability/hand-icap and quality of life measures. A u d i o l o g y40:208–214.

Ricketts T, Henry P, Gnewikow D. (2003) Full timedirectional versus user selectable microphone modesin hearing aids. Ear Hear 24:424–439.

Riko K, Pichora-Fuller MK, Alberti PW. (1986) Clinicalevaluation of a two-channel amplitude compressionhearing aid. Laryngoscope 96:1226–1230.

Page 24: ASystematic Review of Health-Related Quality of Life and ... · J Am Acad Audiol 18:151–184 (2007) 151 *Department of Communication Sciences and Disorders, University of South Florida,

Ringdahl A, Eriksson-Mangold M, Andersson G. (1998)Psychometric evaluation of the Gothenburg Profilefor measurement of experienced hearing disabilityand handicap: applications with new hearing aid can-didates and experienced hearing aid users. Br J Audiol32:375–385.

Salomon G, Vesterager V, Jagd M. (1988) Age relatedhearing difficulties. I. Hearing impairment, disabil-ity and handicap—a controlled study. A u d i o l o g y27:164–178.

Saunders GH, Forsline A, Fausti SA. (2004) The per-formance-perceptual test and its relationship tounaided reported handicap. Ear Hear 25:117–126.

Schum DJ. (1992) Responses of elderly hearing aidusers on the hearing aid performance inventory. JAm Acad Audiol 3:308–314.

Schum DJ. (1993) Test-retest reliability of a short-ened version of the hearing aid performance inventory.J Am Acad Audiol 4:18–21.

Schum DJ. (1999) Perceived hearing aid benefit inrelation to perceived needs. J Am Acad Audiol10:40–45.

Surr RK, Cord MT, Walden BE. (1998) Long-termversus short-term hearing aid benefit. J Am AcadAudiol 9:165–171.

Surr RK, Kolb JA, Cord MT, Garrus NP. (1999)Tinnitus Handicap Inventory (THI) as a hearing aidoutcome measure. J Am Acad Audiol 10:489–495.

Swan IR, Gatehouse S. (1987) Optimum side for fit-ting a monaural hearing aid 2. Measured benefit. BrJ Audiol 21:67–71.

Taylor KS, Jurma WE. (1997) Patients’ t a s k - o r i e n-tation and perceived benefit of amplification inhearing-impaired elderly persons. Psychol Rep81:735–738.

Wesselkamp M, Margolf-Hackl S, Kiessling J. (2001)Comparison of two digital hearing instrument fittingstrategies. Scand Audiol Suppl 30:73–75.

Journal of the American Academy of A u d i o l o g y/ Volume 18, Number 2, 2007

174

Page 25: ASystematic Review of Health-Related Quality of Life and ... · J Am Acad Audiol 18:151–184 (2007) 151 *Department of Communication Sciences and Disorders, University of South Florida,

Quality of Life and Hearing Aids/Chisolm et al

175

STUDY Participant Outcome Level of Other Population Measure Evidence

Baumfield and Dillon (2001) Y N N

Beamer et al (2000) Y N N

Berninger and Karlsson (1999) Y N N

Bridges and Bentler (1998) Y N Y

Brooks et al (2001) Y N Y

Cacciatore et al (1999) Y Y Y Lacked specificity re: p a rticipants’ characteristics

Cord et al (2000) Y Y N

Cox and Alexander (1995) Y N Y

Cox et al (2003) Y N Y

Crowley and Nabelek (1996) Y Y N

Cunningham et al (2001) Y Y N

Dillon et al (1991) Y Y Y Only figures were provided; no data tables

Dye and Peak (1983) Y Y Y Multivariate analysis of variance precluded extraction of data from relevant outcome measures

Gatehouse (1994) Y Y Y No unaided and aided values were provided forcomparison

Gatehouse (1999) Y Y N

Gatehouse and Noble (2004) Y N N

Haggard et al (1981) Y N Y Outcome measures werecreated from existing tools, measurements not provided

Harless and McConnell (1982) Y Y N

Haskell et al (2002) Y N Y

Horwitz and Turner (1997) Y N Y

Humes (2001) Y Y N

Humes (2003) Y Y N

Humes and Wilson (2003) Y Y N

Humes et al (1997) Y N Y

Humes et al (2002a) Y Y Y Only figures were provided; no data tables

Humes et al (2002) Y N Y

Humes et al (2003) Y Y N

Hutton and Canahl (1985) Y N Y

Jerram and Purdy (1997) Y N N

Appendix B. List of Studies That Were Excluded from This Systematic Review andReasons for Their Exclusion Based on A Priori Criteria of Participant Population,

Outcome Measure, Level of Evidence, and Presence of Study Flaws

Page 26: ASystematic Review of Health-Related Quality of Life and ... · J Am Acad Audiol 18:151–184 (2007) 151 *Department of Communication Sciences and Disorders, University of South Florida,

Jerram and Purdy (2001) Y Y N

Kochkin (1997) Y Y N

Kochkin (1998) Y N N

MacKenzie and Browning (1991) Y N Y

Mäki-Torkko et al (2001) N N N

Malinoff and Weinstein (1989b) Y Y Y Duplicate publication

Marttila and Jauhiainen (1996) Y N Y

McCarthy (1996) N N N

Mulrow et al (1992b) Y N N

Newman and Weinstein (1989) Y Y N

Newman et al (1991) Y Y N

Norman et al (1994) Y N Y

Owens and Fujikawa (1980) Y N Y

Palmer et al (1999) Y N N

Parent et al (1997) Y N N

Parving et al (2001) Y N N Purpose was to see if participants could complete the questionnaires

Ricketts et al (2003) Y N Y

Riko et al (1986) Y N Y

Ringdahl et al (1998) N Y N

Salomon et al (1988) Y N Y

Saunders et al (2004) Y Y N

Schum (1992) Y N N

Schum (1993) Y N N

Schum (1999) Y N N

Surr et al (1998) Y N Y

Surr et al (1999) Y N Y

Swan and Gatehouse (1987) N N N

Taylor and Jurma (1997) Y Y N

Wesselkamp et al (2001) N N N

Yueh et al (2001) Y Y Y Study was used in the qualitative analyses, but reporting methods precluded extraction of data for meta-analysis

Journal of the American Academy of A u d i o l o g y/ Volume 18, Number 2, 2007

176

Page 27: ASystematic Review of Health-Related Quality of Life and ... · J Am Acad Audiol 18:151–184 (2007) 151 *Department of Communication Sciences and Disorders, University of South Florida,

Quality of Life and Hearing Aids/Chisolm et al

177

Study Design Participants in Studies Outcome Measure(s) Notes Included in the and ResultsSystematic Review with Meta-Analysis

Abrams et al Quasi-experimental, (1992) parallel group

Chmiel and Overall study designJerger (1996) was quasi-experimental

Data used here were from a pretest/posttest component

Appendix C. Characteristics of Studies That Were Included in This Systematic Review

HHIE

• HA alone group pre-post test differencea

• Control group pre-post test differenceb

A 3rd group, whose data werenot considered, receivedt reatment by HAs plus a 3-weekgroup aural rehabilitationprogram. Random assignmentwas made to two treatmentgroups, but not to the controlgroup.

Control group was advisedthat an HA would bebeneficial; however, they werenot eligible for HAs throughthe VA.

N = 20 veterans whowere new HA users intwo groups

HA alone group:11 MalesM age = 71.3 yr. R = 63–82 yr.

PTA (in dB HL) for 2000, 3000, 4000, and 8000 HzM (SD) = Rt: 63.4 (14.5) L: 53.8 (11.8)

Control group:9 Males M age = 63.9 yr. R = 55–71 yr.

PTA (in dB HL) for 2000, 3000, 4000, and 8000 HzM (SD) = Rt: 46.8 (8.0)L: 43.9 (7.4)

HAs received at nocost

N = 63 new HA users

42 had normalperformance on theDichotic SentenceIdentification task 29 Males13 FemalesM age = 72.4 yr.SD = 5.6 yr.PTA (in dB HL) for 500,1000, and 2000 HzM (SD) =Rt: 32.4 (11.5)L: 35.1 (10.2)PTA (in dB HL) for1000, 2000, and 3000 Hz M (SD) =Rt: 44.8 (10.3)L: 47.3 (9.3)

21 had abnormalperformance on theDichotic SentenceIdentification task13 Males8 FemalesM age = 70.3 yr.

HHIE

• Within n = 63 Ss, pre-post test differencea

The Dichotic SentenceIdentification (Fifer et al, 1983)task assesses central auditoryprocessing.

Study was designed to assesseffect of central auditoryprocessing disorder on self-reported HA outcomes using abetween-groups design.

For this review, only the within-subjects component, wherepre- and post-HHIE scoreswere obtained for baseline andpost-HA use, was considered.

The HHIE was alsoadministered to significantothers.

Page 28: ASystematic Review of Health-Related Quality of Life and ... · J Am Acad Audiol 18:151–184 (2007) 151 *Department of Communication Sciences and Disorders, University of South Florida,

Dillon et al Overall study design (1997) was correlational

Data used here were from a pre-post test component

Humes et al Pre-post test(2001)

Jerger Overall study et al (1996) design was

quasi-experimental

Data used here were from a pre-post test component

Journal of the American Academy of A u d i o l o g y/ Volume 18, Number 2, 2007

178

SD = 7.1 yr.PTA (in dB HL) for 500,1000, and 2000 HzM (SD) =Rt: 33.1 (16.2)L: 37.7 (13.7)PTA (in dB HL) for 1000,2000, and 3000 HzM (SD) =Rt: 46.4 (13.1)L: 50.7 (13.8)

HAs worn as part of atrial period

N = 98 new HA users

Gender distribution notgiven

M age = 71 yr.Inter-quartile R =

67–75 yr.PTA (in dB HL) for 500,1000, and 2000 HzR = 10–20 dB HL and70–80 dB HL; Mode =30–40 dB HL

HAs received at no cost

HHIE

• Pre-post test differencea

Additional subjectivemeasures wereadministered.

Study Design Participants in Studies Outcome Measure(s) Notes Included in the and ResultsSystematic Review with Meta-Analysis

N = 173 105 new and 68experienced HA users

118 Males55 Females

M age = 73.1 yr.SD = 6.5 yr.

HAs purchased

HHIE

• Pre-post test differencea

Additional subjectivemeasures wereadministered.

N = 80 new HA users50 Males30 Females

M age = 74.3 yr.R = 60–96 yr.Better ear:PTA (in dB HL) for 500,1000, and 2000 HzM (R) = 37.4 (16–76)PTA (in dB HL) for 1000,2000, and 3000 HzM (R) = 36.8 (18–67)

HA worn as part of a trial period

HHIE

• Statistically significant difference for all aided relative to unaided conditions, including the conventional HA conditiona

Crossover design wasused to examineoutcomes as a function of:conventional HAs,assistive listeningdevices, and acombination of the two.Only data fromconventional HAs wereconsidered.

A second group of N =100 experienced HAusers was also examined,but no unaided data wereavailable.

Additional subjectivemeasures wereadministered.

Page 29: ASystematic Review of Health-Related Quality of Life and ... · J Am Acad Audiol 18:151–184 (2007) 151 *Department of Communication Sciences and Disorders, University of South Florida,

Quality of Life and Hearing Aids/Chisolm et al

179

Joore et al Pre-post test(2003)

Joore et al Pre-post test(2002)

Malinoff and Pre-post testWeinstein (1989)

Study Design Participants in Studies Outcome Measure(s) Notes Included in the and ResultsSystematic Review with Meta-Analysis

N = 81 new HA users44 Males37 Females

M age = 68 yr.R = 28–95 yr.SD = 11.5 yr.Better ear:PTA (in dB HL) for 1000,2000, and 4000 HzM (SD) = 47.4 (9.9)R = 27–80

HAs received at no cost

EuroQoL EQ-5D andEQ-VAS

Pre-post differencesMobilityb• Self-careb• Daily activitiesb• Pain/Complaintsb• Anxiety/

D e p re s s i o na• EQ-VASb

ADPI-VAS

Pre-post differences• ADPI-VASa

MOS SF-36 SocialFunctioning Scale

Pre-post differences• EQ-5D

• MOS SF-36 Social Functioninga

N = 98 new HA users53 Males45 FemalesM age = 67 yr.R = 28–95 yr.Better ear:PTA (in dB HL) for 1000,2000, and 4000 HzM (SD) = 46 (10)

HAs received at no cost

EQ-5D and EQ-VAS

Pre-post testdifferences

EQ-5D

• Mobilityb• Self-careb• Daily activitiesb• Pain/Complaintsb• Anxiety/

Depressiona

• EQ-VASb

ADPI-VAS

Pre-post test differences

• ADPI-VASa

The ADPI consists of 5additional questions thatassess auditory disability,which were not included inthis review.

The study was designed toassess “response shift” ofbaseline scores. Inaddition to baseline andpost-HA use outcomes,individuals were asked torate their unaidedperformance after HA use(Then-test). Only baselineand HA data were used.

N = 45 new HA users

Gender distribution wasnot reported.

M age = 69.55 yr.R = 55–90 yr.Fitted ear:PTA (in dB HL) for 500,1000, and 2000 HzM (SD) = 45.13 (14.4)R = 15–88

No statement about howHAs were received

HHIE

• Pre-post test differencea

Page 30: ASystematic Review of Health-Related Quality of Life and ... · J Am Acad Audiol 18:151–184 (2007) 151 *Department of Communication Sciences and Disorders, University of South Florida,

Mulrow et al RCT(1990)

Mulrow et al Pre-post test(1992a)

Journal of the American Academy of A u d i o l o g y/ Volume 18, Number 2, 2007

180

Study Design Participants in Studies Outcome Measure(s) Notes Included in the and ResultsSystematic Review with Meta-Analysis

N = 193 veterans whowere new HA users in two groups

Immediate treatmentgroup (ITG)95 Males

M age = 73 yr.SD = 7 yr.Better ear:PTA (in dB HL) for 1000,2000, and 4000 HzM (SD) = 53(10)

Delayed treatmentcontrol group (DTG)98 Males1 Female

M age = 71 yr.SD = 5 yr.Better ear:PTA (in dB HL) for 1000,2000, and 4000 HzM (SD) = 51 (8)

HAs received at no cost

HHIE

• ITG change score > DTG change scorea

QDS

• ITG change score > DTG change scorea

SPMSQ

• ITG change score > DTG change scorea

GDS

• ITG change score > DTG change scorea

SELF

• ITG change scorebDTG change scoreb

N = 192 veterans whowere new HA users

191 Males1 Female

M age = 72 yr.SD = 6 yr.Better ear:PTA (in dB HL) for 1000,2000, and 4000 HzM = 52

HAs received at no cost

HHIE

• Pre vs. all post change at4 monthsa8 monthsa12 monthsa

QDS

• Pre vs. all post change at4 monthsa8 monthsa12 monthsa

SPMSQ

• Pre vs. all post change at4 monthsa8 monthsb12 monthsb

GDS

• Pre vs. all post change at4 monthsa8 monthsa12 monthsa

Follow-up of Mulrow et al(1990) assessing outcomesat 4, 8, and 12 mo.

Page 31: ASystematic Review of Health-Related Quality of Life and ... · J Am Acad Audiol 18:151–184 (2007) 151 *Department of Communication Sciences and Disorders, University of South Florida,

Quality of Life and Hearing Aids/Chisolm et al

181

Newman et al Overall study design (1993) was quasi-

experimental

Data used here were from a pre-post test component

Newman and Pre-post testWeinstein(1988)

Study Design Participants in Studies Outcome Measure(s) Notes Included in the and ResultsSystematic Review with Meta-Analysis

N = 52 new HA users intwo groups

26 uninsured 11 Males15 Females

M age = 74.1 yr.R = 65–85 yr.SD = 5.7 yr.

Fitted/better ear:PTA (in dB HL) for 500,1000, and 2000 HzM (SD) = 34.2 (10)PTA (in dB HL) for 500,1000, and 4000 HzM (SD) = 44.3 (23)

26 insured13 Males13 Females

M age = 75.2 yr.R = 65–86 yr.SD = 6.0 yr.

Fitted/better ear:PTA (in dB HL) for 500,1000, and 2000 HzM (SD) = 35.6 (10)

PTA (in dB HL) for 500,1000, and 4000 Hz

M (SD) = 47.6 (24)

1/2 received HAs viainsurance1/2 paid for HAs

HHIE

• Pre-post test differencea

Study was designed toassess effect of insurancefor HA status on outcomes.This question wasaddressed with a between-groups design. Only within-subject pre–and post–HAfitting data were consideredfor this review.

N = 18 experienced,male HA users

M age = 70.5 yr.R = 66–84 yr.

Better ear:PTA (in dB HL) for 500,1000, and 2000 HzM (SD) = 43.1 (20.5)Poorer ear:M (SD) = 56.6 (24.2)

HAs received at no cost

HHIE

• Pre-post test differencea

The HHIE was alsoadministered to significantothers.

Page 32: ASystematic Review of Health-Related Quality of Life and ... · J Am Acad Audiol 18:151–184 (2007) 151 *Department of Communication Sciences and Disorders, University of South Florida,

Primeau (1997) Overall study design was quasi-experimental

Data used here were from a pre-post test component

Stark and Pre-post testHickson (2004)

Taylor (1993) Pre-post test

Journal of the American Academy of A u d i o l o g y/ Volume 18, Number 2, 2007

182

Study Design Participants in Studies Outcome Measure(s) Notes Included in the and ResultsSystematic Review with Meta-Analysis

N = 233 veterans; newand experienced HAusers

227 Male6 Female

R age = 27–97 yr.

139 older58 new HA users and

81 experienced HAusersM age: 74 yr.SD = 6 yr.

94 younger 54 new HA users and 40 experienced HA

usersM age = 54 yr.SD = 14.5 yr.

Bilateral: PTA (in dB) for 250, 500,1000, 2000, 4000, and8000 HzOlder M (SD) = 54 (14)YoungerM (SD) = 44 (13)

HAs received at no cost

HHIE

• Pre-post test differencea

HHIA

• Pre-post test differencea

Study included conductiveand sensorineural hearinglosses. Only sensorineuraldata were included for thisreview.

Number of older adults withsensorineural hearing losswas 81 (HHIA) and 123(HHIE).

N = 93 new HA users

76 Males17 Females

M age = 71.7 yr.R = 47–90 yr.SD = 8.6 yr.

PTA (in dB HL) for 500,1000, and 2000 Hz R = <25–55

HAs received at no cost

HHIE• Pre-post test

differencea

MOS SF-36P re-post test diff e re n c e s• Physical functionb• Role-Physicalb• Bodily Painb• General Healthb• Vitalityc• Social Functioningc• Role-Emotionalb• Mental Healthb

Significant others wereadministered a modifiedversion of the QDS and theMOS SF-36.

N = 58 new HA users39 Males19 Females

M age = 72.1 yr.R = 65–81 yr.SD = 5.2 yr.Fitted/better ear:PTA (in dB HL) for 2000,3000, and 4000 HzM (SD) = 34.5 (7.7)R = 25–55

HAs purchased

HHIE

• Pre-post test differencea

Page 33: ASystematic Review of Health-Related Quality of Life and ... · J Am Acad Audiol 18:151–184 (2007) 151 *Department of Communication Sciences and Disorders, University of South Florida,

Quality of Life and Hearing Aids/Chisolm et al

183

Yueh et al RCT(2001)

This study was used in the qualitative analysis, but not the quantitative analysis.

Note: HA = hearing aid; L = left ear; M = mean; N = number; R = range; Rt = right ear; Ss = subjects; VA = Veterans Administration.

aStatistically significant increase in HRQoL. bNo significant change in HRQoL as a function of hearing aid use. cStatistically significant decrease in HRQoL.

Study Design Participants in Studies Outcome Measure(s) Notes Included in the and ResultsSystematic Review with Meta-Analysis

N = 60 male veteranswho were all new HAusers in four groups

Control group15 Males

M age = 67 yr.R = 52–85 yr.PTA (in dB HL)M (SD) = Rt: 32.8 (3.7)L: 33.3 (5.0)

Assistive ListeningDevice (ALD) group15 MalesM age = 66.6 yr.R = 53–79 yr.PTA (dB HL)M (SD) = Rt: 32.6 (5.6)L: 32.4 (4.7)

Standard HA group 15 MalesM age = 72.1 yr.R = 53–82 yr.PTA (in dB HL)M (SD) = Rt: 34.6 (5.8)L: 33.0 (6.1)

Programmable HA group 16 MalesM age = 68.5 yr.R = 50–86 yr.PTA (in dB HL)M (SD) = Rt: 31.5 (7.2)L: 31.0 (6.9)

HAs received at no cost

HHIE

Pre-post testdifferencea

Additional self-reportmeasures wereadministered, but onlythe HHIE was consideredan HRQoL measure forthis review.Diary data wereconsidered HRQoLqualitative data.

Data from all groupsexcept ALD were relevantfor the present analysis.

Frequencies for PTA werenot specified.