Guidelines for development of patient-reported outcome measures:

As PROM utilization has increased in importance, so too has the need to follow stringent guidelines to develop these instruments. The Patient-Reported Outcomes Measurement Information System (PROMIS) and COnsensus-based Standards for the selection of health status Measurement INstruments (COSMIN) have established standards that aim to improve the quality of PROMs used to measure clinical and research outcomes(Mokkink et al., 2010; PROMIS). In general, these standards support a mixed methods research design that include: development of a construct to be measured, systematic review of prior work done on the topic, inclusion of a sample of affected patients (key-informant interviews or focus groups), cognitive interviews with the sample patients, use of modern psychometric techniques for analyses, and reliability and validity testing to evaluate the final instruments. Both standards have the ultimate aim of developing instruments that accurately portray the values of a population and stratify patients along the ability continuum of the construct of interest.

These methodologies have rarely been applied to adults with hearing loss and never to CI users prior to the current work (McRackan, Hand, et al., 2018; McRackan, Hand, Velozo, Dubno, et al., 2019). The importance of doing so is three-fold. First, incorporating stakeholder engagement in PROM development provides face and content validity by ensuring topics that affect QOL of the population of interest are included in hearing-related and CI-specific PROMs (McRackan et al., 2017). Second, cognitive interviewing ensures the items included in the final instruments are relevant to the lives of CI users and the interpretation of the item responses aligns with the intention of the research team. Third, as discussed later, the use of modern psychometric analyses has the potential to improve the measurement properties of PROMs.

Item response theory (IRT) is the core of these modern psychometric analyses used to develop PROMs and has several advantages over classical test theory (the previous standard for PROM development). First, classical test theory is grounded on observed and true scores as opposed to IRT, which focuses on the measurement of an underlying trait—referred to as person ability or person measure. As such, instruments developed with classical test theory are sample-dependent as subjects will have higher true scores on easier tests and lower true scores on more difficulty tests. In contrast, instruments developed using IRT remain sample and test independent (Prieto et al., 2003).

A second advantage of IRT is the focus on item-level, rather than test-level, psychometrics (as in classical test theory). The data provided by IRT analysis on each individual item determines its measurement characteristics and utility for inclusion in subsequent instruments. For example, IRT analysis evaluates each item for ceiling and floor effects, identifies fit to the hierarchical model, matches individual item difficulty level to person ability level, and ensures that the items cover the ability range of the population of interest. These results create an item bank that measures and differentiates individuals across the range of ability levels and serves as the source for items to be used for subsequent PROMs including short form, profile, and computerized adaptive testing (CAT) instruments. Given that the psychometric properties of each item are established, researchers can then select items for each instrument based on their highest discrimination across the ability range and best match between item difficulty and subject ability. This process results in optimized instruments with maximized capacity to differentiate individuals across a greater range of the latent trait—termed precision(Rose et al., 2008).

The third advantage relative to classical test theory is based on the more strict assumptions that must be met before IRT analysis is performed, including (Reeve et al., 2007): (1) items only contribute to one domain of QOL (unidimensionality), (2) responses to each item are unrelated to responses to other items (local independence), and (3) items fit the IRT measurement model (item fit). Confirmatory factor analysis (CFA) is used to confirm unidimensionality and local independence. Items are eliminated if they do not significantly contribute to the unidimensional construct captured by the other items in a domain, or if responses to the item are dependent upon responses to other items in the pool. In addition, item fit to the IRT model, such as infit and outfit, are examined to ensure that the included items sufficiently measure the construct of interest for individuals at ability levels close to and far from the item difficulty.

Legacy instruments

The use of hearing-related and CI-specific PROMs is important in for assessments of CI users as data suggest that patients report two times greater improvement when using these instruments rather than generic health-related QOL instruments (McRackan, Bauschard, Hatch, Franko-Tobin, Droghini, Nguyen, et al., 2018; McRackan, Bauschard, Hatch, Franko-Tobin, Droghini, Velozo, et al., 2018; McRackan, Fabie, et al., 2019). However, the legacy hearing-related and CI-specific PROMs most often used to assess CI users were not developed using the stringent PROMIS and COSMIN guidelines to establish face and construct validity. In fact, commonly used instruments, such as the Hearing Handicap Inventory for Adults/Elderly (HHIA/E) (Newman et al., 1990; Ventry & Weinstein, 1982), Speech, Spatial and Quality of Hearing Scale (SSQ) (Gatehouse & Noble, 2004), and the Abbreviated Profile of Hearing Aid Benefit (APHAB) (Cox & Alexander, 1995), were developed and validated in individuals with mild to moderate hearing loss and hearing aid users. Given that CI users and those with more severe hearing loss were excluded from the development process, researchers and clinicians using these PROMs cannot be confident that the results accurately and reliably reflect the themes and constructs for adult CI users (Capretta & Moberly, 2016; Park et al., 2011; Vermeire et al., 2006; Vermeire et al., 2005).

The Nijmegen Cochlear Implant Questionnaire (NCIQ) is by far the most commonly used CI-specific PROM (Hinderink et al., 2000; McRackan, Bauschard, Hatch, Franko-Tobin, Droghini, Nguyen, et al., 2018). As summarized in a recent meta-analysis (McRackan, Bauschard, Hatch, Franko-Tobin, Droghini, Nguyen, et al., 2018), NCIQ total score and all subdomain scores improve after cochlear implantation. The NCIQ consists of 60 items with 3 domains (physical, psychological, and social). The domains are further divided into 3 subdomains: physical includes basic sound perception, advanced sound perception, and speech production; psychological includes self-esteem; social includes activity limitations and social interactions. The NCIQ includes items and domains selected by expert opinion, rather than from information provided by CI users. As such, the NCIQ does not contain certain domains and themes that have been shown to be of value to CI users in subsequent qualitative work (Hughes et al., 2018; McRackan et al., 2017). Again, this demonstrates the value of directly soliciting input from CI users to ensure that a QOL instrument adequately captures concepts that are important to members of the target population. Such stakeholder engagement is now considered standard procedure for PROM development (Mokkink et al., 2010; PROMIS). In addition, the NCIQ was only validated and tested in Dutch on 91 participants (including 46 controls) from a single clinical site in the Netherlands without cross-cultural adaptation(Hall et al., 2018) in other populations. Finally, as detailed earlier, more rigorous psychometric testing for PROM development has become standard since the NCIQ was developed.

The establishment of PROMIS and COSMIN guidelines for best-practices in PROM development provides a new opportunity for the field of hearing science to critically re-examine existing instruments and develop novel PROMs following these rigorous guidelines.

Development of the CIQOL instruments

The stages of CIQOL instrument development have been previously described in detail but are outlined here (McRackan, Bauschard, Hatch, Franko-Tobin, Droghini, Nguyen, et al., 2018; McRackan, Bauschard, Hatch, Franko-Tobin, Droghini, Velozo, et al., 2018; McRackan, Hand, et al., 2018; McRackan et al., 2017). In general, this process included three main stages: (1) creation of the new Cochlear Implant Quality of Life (CIQOL) item pools; (2) psychometric analysis of the CIQOL pools to create the final CIQOL item banks; (3) selection of items for the CIQOL instruments based on the results of IRT analysis of the item bank.

Subjects and Recruitment

Subjects were recruited through the CIQOL Development Consortium, which consists of 30 institutions and was established to recruit a large sample of adult CI users representative of the broader adult CI population. Recruitment flyers were distributed electronically and on paper to CI users through these CI centers. Potential subjects emailed our research team in order to be enrolled. Subjects were required to: (1) be between 18–89 years of age (as individuals >89 years of age are considered a special population), (2) have used a CI for over one year, and (3) not have received a CI for single sided deafness. The subjects recruited for the current study were an independent cohort from those recruited previously to develop the CIQOL item bank.

Measures

The questionnaire consisted of three major sections: (1) subject demographics (completed only at Time 1), (2) the CIQOL-35 Profile (which includes all items in the CIQOL-10 Global), and (3) legacy instruments - the NCIQ, and HUI-3. Subject demographics included: age, sex, marital status, whether subjects had children (<18 years old) in the household, household income, education level, geographic region, residential area description, and hearing/CI history. Geographic regions were determined by the US Census Bureau definitions (Bureau, 2010) and subjects self-identified the developed area where they lived as urban, suburban, or rural. Regarding hearing and CI history, duration of CI use was calculated based on the date subjects had their first (or only) CI activated. Subject’s listening modality was defined as the hearing device configuration they used on a routine basis (bilateral CIs, unilateral CI with or without contralateral hearing aid). Subjects also obtained their most recent best aided speech recognition scores from their audiologist and entered them into the questionnaire. These scores could include word scores (Consonant-Nucleus-Consonant [CNC]), sentence scores in quiet (Hearing in Noise Test [HINT]; AzBio quiet), or sentence scores in noise (AzBio +10 dB signal-to-noise ratio [SNR]), as these are components of the minimum reporting standards (Adunka et al., 2018). Subjects were not excluded if they could not obtain speech recognition scores.

For the CIQOL-35 Profile, subjects respond to each item using a 5-point scale, where higher scores indicate better QOL. Scores are derived by calculating a sum of subject responses and using score conversion tables that were produced using IRT (Cochlear Implant Quality of Life Research Program, 2019). Subjects also completed the two legacy instruments - the NCIQ and HUI-3. Each item in the NCIQ is rated on a 5-point scale and item-level responses are averaged to yield scores for each NCIQ subdomain, domain, and total score. Scores were calculated based on the corrected table published after the initial NCIQ (“Corrigendum,” 2017). The HUI-3 consists of 8 items, each corresponding to an attribute of health (vision, hearing, speech, ambulation, dexterity, emotion, cognition, and pain). Respondents select their health level for each attribute using a scale ranging from 1–5 (for speech, emotion, and pain) or 1–6 (for all other attributes), where higher scores indicate poorer health related QOL. Scores for each individual attribute are derived using a weighting system for the subject’s item-level response (single-attribute utility function) and a HUI-3 total score is derived from the multi-attribute utility function (Horsman et al., 2003). Subjects were randomized regarding the order in which they completed the CIQOL, NCIQ, and HUI-3 instruments.

Construct validity

Confirmatory Factor Analysis

The CFA results are shown in Table V. Most CIQOL-35 Profile domains and CIQOL-10 Global demonstrated good model fit. RMSEA indicated poor model fit for all CIQOl-35 domains except communication. However, this may be because RMSEA is a less reliable indicator of fit (Kenny et al., 2015) in models with few degrees of freedom. The other fit indicators included accommodate fewer degrees of freedom. In addition, the entertainment domain did not meet our threshold for acceptable SRMR. However, we proceeded with Rasch analysis given that the SRMR (0.09) was very close to our cut-off threshold of 0.08, previous work demonstrated adequate fit (McRackan, Hand, Velozo, Dubno, et al., 2019), and the two other fit indices examined were adequate. All CIQOL-35 Profile domains and CIQOL-10 Global met a-priori CFI and TLI model fit criteria. In addition, all items had standardized factor loading of ≥│0.32│ on their respective domains.

Table V:

Confirmatory factor analysis results for the CIQOL-35 Profile, CIQOL-10 Global, and NCIQ.

Instrument	SRMR	RMSEA	CFI	TLI	% of items with standardized factor loadings ≥|0.32|
CIQOL-35 Profile
Communication	0.04	0.06	>0.99	>0.99	100
Emotional	0.05	0.10	>0.99	0.99	100
Entertainment	0.09	0.27	0.99	0.99	100
Environmental	0.08	0.20	0.99	0.98	100
Listening effort	0.05	0.10	0.99	0.99	100
Social	0.05	0.11	>0.99	0.99	100
CIQOL-10 Global	0.04	0.06	>0.99	>0.99	100
NCIQ
Physical	0.08	0.11	0.69	0.67	93
Basic sound perception	0.03	0.06	0.97	0.97	100
Advanced sound perception	0.07	0.15	0.82	0.77	100
Speech production	0.09	0.16	0.74	0.67	100
Psychological	0.06	0.10	0.87	0.84	90
Self-esteem	0.06	0.10	0.87	0.84	90
Social	0.05	0.08	0.89	0.88	95
Activity limitation	0.03	0.06	0.97	0.96	100
Social interaction	0.04	0.08	0.94	0.99	90
Total	0.08	0.08	0.64	0.63	92

Open in a separate window

Bolded text designates values outside of our a-priori established model fit criteria as described in the methods section. SRMR: standardized root mean square residual; RMSEA: root mean square error of approximation; CFI: comparative fit index; TLI: Tucker-Lewis index

Apart from the basic sound perception and activity limitation subdomains, all NCIQ domains and subdomains had multiple indicators of poor model fit (bolded in Table V). Given that unidimensional structure is required for IRT analysis, further psychometric analysis with IRT could only be performed on the basic sound perception and activity limitation subdomains.

Item Response Theory

All CIQOL-35 Profile domains, the CIQOL-10 Global, and NCIQ subdomains that were analyzed with IRT met our three criteria for rating scale appropriateness (i.e., monotonicity, > 10 observations per rating scale category, and outfit mean square < 2.0). Additionally, all analyses revealed: an absence of substantial ceiling or floor effect (<15% of subjects), good to high person reliability (i.e., ≥0.80), acceptable numbers (>3) of statistically distinct person strata that can be reliably differentiated, and a relatively high Cronbach’s alpha ≥0.80. Table VI provides a summary of the IRT analysis results.

Table VI:

Summary of item response theory (IRT) analysis results for the CIQOL-35 Profile, CIQOL-10 Global, and NCIQ subdomains that met criteria for unidimensionality.

Instrument	Item Misfit	Person Misfit (%)	Ceiling / Floor (% of Subjects)	Person Reliability	Person Strata	Mean Person Measure (logits)	Cronbach’s alpha
CIQOL-35 Profile
Communication	1	7.78	0.00 / 0.00	0.91	4.56	0.22	0.92
Emotional	0	7.78	2.69 / 0.00	0.82	3.16	1.58	0.84
Entertainment	1	7.49	6.89 / 2.99	0.86	3.60	0.63	0.91
Environmental	0	6.89	3.29 / 0.30	0.84	3.36	1.83	0.87
Listening effort	0	5.39	0.00 / 0.60	0.83	3.32	−1.04	0.84
Social	1	7.49	8.08 / 0.00	0.83	3.28	2.39	0.88
CIQOL-10 Global	0	6.59	0.00 / 0.00	0.87	3.78	0.55	0.88
NCIQ
Physical
Basic sound perception	1	7.49	8.08 / 0.00	0.86	3.61	0.92	0.89
Advanced sound perception	—	—	1.19 / 0.30	—	—	—	—
Speech production	—	—	15.27 / 0.00	—	—	—	—
Psychological	—	—	0.00 / 0.00	—	—	—	—
Self-esteem	—	—	0.00 / 0.00	—	—	—	—
Social	—	—	0.00 / 0.00	—	—	—	—
Activity limitation	0	6.59	0.00 / 0.00	0.84	3.42	1.70	0.90
Social interaction	—	—	0.00 / 0.00	—	—	—	—
Total	—	—	0.00 / 0.00	—	—	—	—

Open in a separate window

Bolded text designates values outside of our a-priori established criteria as described in the methods section. Dashes represent the domains and subdomains where IRT was not performed due to poor model fit.

The CIQOL-35 Profile communication, entertainment, and social domains, as well as the NCIQ basic sound perception and activity limitation subdomains, each had one item that demonstrated model misfit. Four CIQOL-35 Profile domains and both NCIQ sub-domains had mean person measures > |1.0|, indicating a slight mismatch between subject ability and item difficulties. However, these values are acceptable given low ceiling and floor effects for these domains. For example, the mean person measure for the CIQOL-35 Profile social domain revealed that the average subject ability was 2.4 logits higher than the mean item difficulty. However, given that only 8.1% of subjects showed a ceiling effect for this domain, it appears that the full range of ability is being measured in the adult CI population.

While IRT analysis could not be completed on the additional NCIQ domains and subdomains, ceiling and floor effects were calculated (Table VI). No substantial ceiling or floor effects were observed except for the NCIQ speech production subdomain where 15.3% of subjects reported the highest possible score.

Convergent validity

Pearson correlations between the CIQOL-35 Profile domains and CIQOL-10 Global scores with conceptually similar components of the NCIQ and HUI-3 are shown in Table VII and Figure 2. Specifically, Table VII displays correlations between the NCIQ/HUI-3 composite scores and all CIQOL domains and conceptually similar NCIQ and CIQOL domains. Figure 2 compares conceptually similar NCIQ subdomains and CIQOL domains. Scores on the CIQOL-35 communication domain were strongly correlated with the NCIQ basic sound perception, advanced sound perception, physical function, and total score (r=0.70–0.83) and weakly correlated with speech production (r=0.42; Figure 2). The CIQOL-35 emotional domain and social domain scores were strongly correlated with scores on the NCIQ psychological domain (r=0.80) and social function domain (r=0.72), respectively. The CIQOL-10 Global score was strongly correlated with the NCIQ total score (r=0.85).

Open in a separate window

Figure 2:

Convergent validity between comparable CIQOL domains and NCIQ subdomains and HUI-3 health attributes. Correlation coefficients for each comparison are provided with 95% confidence intervals. Solid line represents the line of best fit. BSP: basic sound perception; ASP: advanced sound perception

Table VII:

Convergent validation of CIQOL instruments with legacy measures where conceptually similar domains exist (Table I).

	CIQOL-35 Profile						CIQOL-10 Global
Instrument and domain	Communication r (95% CI)	Emotional r (95% CI)	Entertainment r (95% CI)	Environmental r (95% CI)	Listening Effort r (95% CI)	Social r (95% CI)	r (95% CI)
NCIQ
Physical	0.78 (0.74, 0.82)	—	—	—	—	—	—
Psychological	—	0.80 (0.76, 0.84)	—	—	—	—	—
Social	—	—	—	—	—	0.72 (0.66, 0.76)	—
Total	0.81 (0.77, 0.84)	0.73 (0.67, 0.77)	0.68 (0.62, 0.74)	0.73 (0.67, 0.77)	0.73 (0.68, 0.78)	0.77 (0.72, 0.81)	0.85 (0.82, 0.88)
HUI-3 Total	0.31 (0.21, 0.41)	0.23 (0.12, 0.33)	0.27 (0.17, 0.36)	0.29 (0.19, 0.39)	0.24 (0.14, 0.34)	0.2 (0.09, 0.3)	0.29 (0.19, 0.38)

Open in a separate window

Values represent Pearson correlation coefficients with strong relationships in bold. Subdomain data are displayed in Figure 2. Dashes denote correlations not performed as the respective domains/subdomains had no hypothesized association based on the constructs they purport to measure. NCIQ physical domain includes basic sound perception, advanced sound perception, and speech production; psychological includes self-esteem; social includes activity limitations and social interactions.

Overall, HUI-3 scores were weakly correlated with the CIQOL-35 Profile scores and CIQOL-10 Global scores. Specifically, the CIQOL-35 Profile communication domain demonstrated low correlation with the HUI-3 speech and hearing scores (r=0.17–0.32; Figure 2). All CIQOL-35 Profile domains and the CIQOL-10 Global were weakly correlated with HUI-3 total scores (r=0.20–0.31; Table VII).

Use of HUI-3 in adult CI users

Generic health-related QOL instruments, such as the HUI-3, SF-36 and EQ-5D are routinely used to determine the health utility and health economic benefit from a medical intervention. The HUI-3 was selected as a comparator legacy measure for the current study based on the frequency of use (Arnoldner et al., 2014; Damen et al., 2007; McRackan, Bauschard, Hatch, Franko-Tobin, Droghini, Velozo, et al., 2018; Palmer et al., 1999; Smulders et al., 2016; van Zon et al., 2017) and data suggesting that it may be valid for adult CI users (Yang et al., 2013). Based on responsiveness data and differentiation of subjects with and without hearing loss, Yang et al.(Yang et al., 2013) reported that the HUI-3 was the best available measure to evaluate generic health-related QOL in individuals with hearing loss, but noted the lack of reliability data for the HUI-3 in the hearing loss population. This may be a reasonable assumption as the HUI-3 is the only such instrument that includes a hearing dimension. However, the wording of the hearing dimension may help explain the poor reliability of the HUI-3 in CI users who participated in our study.

The HUI-3 hearing dimension asks users to identify their health level on a 1–6 scale. Interestingly, levels 1–5 all specifically mention the use of hearing aids and level 6 is simply “Unable to hear at all.” There is no mention of CIs or other assistive listening devices, which may lead to confusion and different interpretations in CI users. For example, some users may consider their CI a hearing aid while others do not. In addition, many CI users communicate with bimodal hearing, that is, a CI in one ear and hearing aid in the other (31.4% in the current study), causing further confusion on these items. The wording of the HUI-3 hearing dimension is logical when measuring QOL in a large sample of individuals with hearing loss, where hearing aid use far outnumbers CI use, but is problematic when applied specifically to CI users. At times, the HUI-3 hearing dimension has been altered to read “hearing aid or cochlear implant,” (Palmer et al., 1999) but it is unclear to what extent this changes the instrument’s measurement properties and alters the health utility index calculation. We opted to maintain the original, validated language for the current study so our results could be more readily compared with others in the literature. Nevertheless, this wording may be responsible for a substantial portion of the low correlation between the HUI-3 hearing dimension and the CIQOL-communication domain.

The wording of the HUI-3 hearing dimension is also likely responsible for the relatively low observed test-retest reliability of this domain in the current study. For example, patients may interpret the item stem differently, as described earlier, on repeated administrations. As the CIQOL-35 Profile, CIQOL-10 Global, and NCIQ demonstrated high test-retest reliability, it is unlikely that changes in functional ability account for the low test-retest reliability of the HUI-3 hearing dimension. Although all instruments/domains demonstrated relatively low correlations with speech recognition scores, the HUI-3 hearing dimension values were even lower, which is consistent with published data (Damen et al., 2007; Kumar et al., 2016). Thus, we do not recommend using the HUI-3 to evaluate health utility benefit in adult CI users and recommend caution in interpreting results from previous studies that have done so.

Given this, researchers are currently limited in the availability of generic health-related QOL measures for use in health economic assessments in adult CI users. Most generic health-related QOL measures include domains such as mobility and bodily pain that are unrelated to cochlear implantation. Thus, these instruments may underestimate the benefits of cochlear implantation, as has been previously reported (McRackan, Bauschard, Hatch, Franko-Tobin, Droghini, Nguyen, et al., 2018; McRackan, Bauschard, Hatch, Franko-Tobin, Droghini, Velozo, et al., 2018; McRackan, Fabie, et al., 2019). Future research is needed to determine which instruments can be used to more accurately estimate benefits of cochlear implantation within the health economic framework.

Implementation of the CIQOL instruments

The current project represents the culmination of a multi-year and multi-institutional study that demonstrates the value of following stringent guidelines and using qualitative to quantitative approaches to develop CI-specific PROMs. From the initial CI focus groups to this validation study, 748 unique adult CI users from around the United States participated in the development of the CIQOL instruments. These results were used to develop instruments that represented the important themes for adult CI users and had superior measurement properties to the currently available PROMs. The results of this study revealed that the CIQOL-35 Profile and CIQOL-10 Global demonstrate strong convergent validity with conceptually similar domains on the NCIQ, yet the CIQOL instruments have stronger evidence of content validity. Additionally, the CIQOL-35 Profile and CIQOL-10 Global demonstrated test-retest reliability in adult CI users similar to the NCIQ and superior to the HUI-3.

Given these results, the CIQOL-35 Profile and CIQOL-10 Global are ready for clinical and research use in adults with CIs. These direct measures complement data obtained from speech recognition outcomes and provide a broader understanding of the benefits of cochlear implantation to an individual’s life. The comprehensive nature of the CIQOL-35 Profile provides assessment of outcomes in domains unrelated to speech recognition ability as measured using traditional tasks. Further, the weak correlation between speech recognition scores and the CIQOL-communication domain suggest that this domain measures a seemingly distinct construct of real-world functional communication abilities. Therefore, the CIQOL instruments can serve as valuable supplements to the traditional speech recognition test battery to evaluate post-CI progress and also the extent to which new CI device technologies, listening modalities, and processing strategies may improve outcomes.

Study limitations

Limitations of the current study are mostly related to the methods used to enroll patients. To make results generalizable, a large sample of adult CI users was recruited from a 30-member CI center consortium. Overall, it appears that this sample may have received more education and had higher income than the general population of the United States. However, it is important to note that the demographics and potential disparities in CI care in the United States have not been thoroughly studied and remain largely unknown. In any study, there is a risk of selection bias where those with strong opinions, positive or negative, may be those most likely to enroll in research studies. We aimed to minimize this effect by enrolling a large patient sample. In addition, we can never know if participants quickly filled out the questionnaires randomly to obtain remuneration. Time to completion data was not obtainable through REDCap. However, such individuals should have been found to misfit the IRT model during analysis and, if large in number, would have affected the reliability data. Neither was detected in the study.

The CIQOL instruments are available for download at: https://education.musc.edu/CIQOL. This work was presented at the 2019 Conference on Implantable Auditory Prostheses (CIAP). The Cochlear Implant Quality of Life Development Consortium consists of the following institutions (and individuals): Columbia University (Justin S. Golub, MD, MS), Duke University Hospital (Erin Blackburn, AuD; Howard Francis, MD; Amy Walker), Eastern Virginia Medical School (Stephanie Moody-Antonio, MD), Georgetown University (Michael Hoa, MD), House Ear Clinic (Eric P. Wilkinson, MD; Dawna Mills, AuD), Johns Hopkins University (John P. Carey, MD), Kaiser Permanente-Los Angeles (Nopawan Vorasubin, MD), Kaiser Permanente-San Diego (Vickie Brunk, AuD), Loyola University Medical Center (Matthew Kirchner, MD), Massachusetts Eye and Ear Infirmary (Kevin Frank, PhD; Elizabeth DesRoche, AuD), Mayo Clinic Rochester (Matthew L. Carlson, MD; Collin L. Driscoll, MD, Medical University of South Carolina (Elizabeth L. Camposeo, AuD; Paul R. Lambert, MD; Ted A. Meyer, MD PhD; Cameron Thomas, BS), New York Eye and Ear Infirmary (Maura Cosetti, MD), The Ohio State University (Aaron C. Moberly, MD), Rush University (Mike Hefferly, PhD; Mark Wiet, MD), Stanford University (Nikolas H. Blevins, MD; Jannine B. Larky, AuD), State University of New York-Downstate (Matthew Hanson, MD), Summit Medical Center (Jed Kwartler, MD), University of Arkansas for Medical Sciences (John Dornhoffer, MD), University of Cincinnati (Ravi N. Samy, MD), University of Colorado (Samuel P. Gubbels, MD), University of Maryland School of Medicine (Ronna P. Herzano, MD, PhD), University of Miami (Michael E. Hoffer, MD; Meredith A. Holcomb AuD; Sandra M. Prentiss, PhD), University of Pennsylvania (Jason Brant, MD), University of Texas Southwestern (Jacob B. Hunter, MD; Brandon Isaacson, MD; J. Walter Kutz, MD), University of Utah (Richard K. Gurgel, MD), Virginia Mason Medical Center (Daniel M. Zeitler, MD), Washington University-Saint Louis (Craig A. Buchman, MD; Jill B. Firszt, PhD); Vanderbilt University (Rene H. Gifford, PhD; David S. Haynes, MD; Robert F. Labadie, MD, PhD).

Funding and conflicts of interest: This research was supported (in part) by a grant from the National Institutes of Health, National Institute on Deafness and Other Communication Disorders, Grant Number K23 DC016911, and a grant from the American Cochlear Implant Alliance. TRM is on the medical advisory board for Envoy Medical.