Abstract
INTRODUCTION:
There are no standard practices for considering sensory impairment in studies measuring cognitive function among older adults. Exclusion of participants with impairments may inaccurately estimate the prevalence of cognitive impairment and dementia.
METHODS:
We surveyed prospective cohort studies measuring cognitive function in older adults, determined the proportion that excluded participants based on sensory impairment and the proportion that assessed each type of sensory impairment, and described the methods of sensory assessment.
RESULTS:
Investigators/staff from 85 (of 192 cohorts) responded; 6 (7%) excluded participants with severe impairment; 80 (94%) measured hearing and/or vision impairment, while 5 (6%) measured neither. Thirty-two (38%) cohorts assessed hearing objectively and 45 (53%) assessed vision objectively.
DISCUSSION:
Findings indicate variation in methods used to assess sensory impairment, with potential implications for resource allocation. To ensure equitable inclusion of study participants, consensus is needed on best practices standardized protocols for assessment and accommodations of sensory impairment.
Keywords: hearing loss, vision loss, sensory impairment, cognitive impairment, dementia, older adults
INTRODUCTION
More than half of all Americans over the age of 60 have hearing or vision impairment and 11.3% of those over the age of 80 have both, often referred to as dual sensory impairment.1,2 These impairments may have important consequences for older adults’ health. Both hearing and vision loss are independently associated with accelerated cognitive decline and increased risk of Alzheimer’s disease and related dementias (ADRD).3–6 However, it is unclear whether these associations are due to causal mechanistic pathways such as increased social isolation and changes in brain structure, or due to bias related to cognitive testing.
Neurocognitive tests are generally administered aurally or on paper, requiring the ability to hear and/or see testing materials. Older adults with sensory impairment may therefore be unable to engage with test materials equitably. In research studies, exclusion of participants with sensory impairment from cognitive testing or from the analysis may occur out of concern for the validity of the results. If participants with sensory impairment cannot hear or see the testing materials, low cognitive test scores may reflect sensory rather than cognitive function; consequently, cognitive impairment in these participants may be overestimated. Depending on the study, this exclusion could impact a large proportion of participants. For example, in a study of cognitive performance in patients with chronic kidney disease, nearly half of the original sample was excluded by investigators due to “visual, motivational or motor difficulties”.7 Alternatively, if there is a causal relationship between sensory impairment and ADRD, exclusion of participants with sensory impairment could have the opposite effect, resulting in an underestimation of the prevalence of ADRD and cognitive impairment. In support of this possibility, we recently found that older adults with hearing loss in the Atherosclerosis Risk in Communities study were less likely to complete cognitive testing than older adults without hearing loss, and that this missing data resulted in an underestimation of the hearing loss-cognitive performance relationship by ~30%.8
Discriminating between the contributions of sensory impairment to testing bias and the contributions of true cognitive impairment will allow us to better target and remediate modifiable pathways. Despite the importance of this distinction, few studies have carefully considered the potential impact of sensory impairment on cognitive testing. Among older adults with documented sensory impairments, there is no consensus across studies on a standardized method used to collect and analyze cognitive testing data in research settings,9 and there have been no formal analyses of research protocols to identify the degree of variation in these methods. Assessing the current landscape of cognitive testing protocols is a necessary first step to informing best practices for the measurement of cognitive function in older adults with sensory impairment.
In the present study, we aimed to describe variations in current methods used to assess and analyze cognitive data, with a focus on how studies consider hearing or vision impairment. To address this aim, we first conducted a systematic review of the literature to identify studies that measure cognitive function in older adults, then administered a survey to staff from each study in order to determine if and how sensory impairment was considered. We hypothesized that there would be significant variation among protocols with respect to the assessment of sensory impairments, exclusions due to sensory impairment, and provision of accommodations.
METHODS
We systematically identified longitudinal cohort studies of older adults that measured cognitive function from peer-reviewed literature. We conducted a literature review using adapted Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) guidelines for the identification and selection of studies. We searched three electronic databases (MEDLINE, Embase, Cochrane Library) for relevant articles using search terms related to cognitive function, prospective study design, and older adults. Records were screened and included if the following criteria were met: prospective study design, mean age of participants ≥60 years, and at least two assessments of cognitive performance. Additionally, we restricted to studies that included at least 200 participants with the goal of eliminating clinic-based samples, as estimates of the prevalence and incidence of ADRD used for public health planning and resource allocation are primarily derived from large population-based studies. We restricted to cohorts that were ongoing or had been completed in the past 5 years, as our goal was to identify and contact current study staff who would be able to complete the survey (Supplemental material – search strategy).
From the studies included after full-text review, we identified a list of unique cohorts. We contacted the principal investigators (PI) or primary staff members for each cohort with a request to participate in a brief online survey. Two rounds of emails were sent to unique cohorts in April-June, 2019 and June-August, 2020 with a short description of the project and a link to the survey. Survey questions included if and how the study assessed hearing and/or vision loss; if and how accommodations for hearing and/or vision loss were provided during cognitive testing; and inclusion/exclusion criteria (Supplemental material – example survey). Accommodations are any additions to the protocol that amplify the visual or auditory stimuli, including but not limited to an amplification device for hearing impairment or reading aloud, using enlarged font, or providing a magnifying glass for vision impairment; use of glasses or hearing aids is not considered an accommodation. Respondents were asked to upload protocol documents describing methods for the collection and analysis of cognitive data, and of hearing and vision if applicable. A follow-up interview was offered to discuss protocol information by telephone if this was preferred to directly uploading the protocol. Reminder emails were sent if no response was received after 7, 14, and 21 days.
We described the number and proportion of cohorts that assessed each type of impairment as well as the method of assessment (broadly categorized as objective or subjective). We cross-tabulated the number and proportion of cohorts that assessed both hearing and vision impairment, only one of the two, or neither, using 1) any method, 2) objective methods, or 3) subjective methods. Examples of objective methods included audiometry, finger rub test, whisper test, and hear screen for hearing10–12; and distance or near acuity, contrast sensitivity, and visual fields for vision.13–17 Subjective methods of assessing hearing or vision included participant self-report and provider-report. We determined the number and proportion of cohorts that used exclusion criteria based on cognitive, vision or hearing impairment during neurocognitive testing. Finally, we described the number and proportion of cohorts that offered accommodations for each type of sensory impairment during neurocognitive testing.
RESULTS
Systematic review
We identified 82,109 records through database searching, 24,517 of which were screened after removing duplicates as well as records not relevant to the study question. A total of 898 records fit the inclusion criteria after the full-text review and 228 unique cohorts were identified from these records. We contacted all cohorts with available contact information (N=192; 84.2%), and 85 (44.3%) provided complete responses to our survey questions (Figure 1).
FIGURE 1.
Inclusion/exclusion criteria for the systematic review
Assessment of sensory impairment
Of the 85 cohorts included in our analysis, 80 (94.1%) measured either hearing (N=77) or vision (N=78) and 5 (5.9%) measured neither. Of the cohorts that measured hearing, 32 (38%) used objective assessments, with audiometry (21 out of 32 cohorts; 66%) and whisper test (7 out of 32 cohorts; 22%) as the most frequently used methods. Seventy-five (88%) cohorts subjectively assessed hearing; 67 used patient-report and 11 used provider-report. Forty-five (47%) cohorts objectively assessed vision, of which nearly 90% used distance or near visual acuity (40 out of 45). Of the 68 (80%) cohorts that subjectively assessed vision, 61 used patient-report and 10 used provider-report (Table 1).
Table 1.
Assessments and accommodations for hearing and vision impairment (N=85)
| Hearing Impairment | Vision Impairment | |
|---|---|---|
| Any Assessment | ||
| Yes | 77 (90.6%) | 78 (91.8%) |
| No | 8 (9.4%) | 7 (8.2%) |
|
| ||
| Objective Assessment | ||
|
| ||
| Tested | ||
| Yes | 32 (37.6%) | 45 (52.9%) |
| No | 52 (61.2%) | 39 (45.9%) |
| Don’t know | 1 (1.2%) | 1 (1.2%) |
|
| ||
| Method (hearing)1 | ||
| Audiometry | 21/32 (65.6%) | -- |
| Finger Rub Test | 3/32 (9.4%) | -- |
| Whisper Test | 7/32 (21.9%) | -- |
| Hear Screen | 1/32 (3.1%) | -- |
| Other | 5/32 (15.6%) | -- |
|
| ||
| Method (vision)1 | ||
| Distance or Near Acuity | -- | 40/45 (88.9%) |
| Contrast Sensitivity | -- | 14/45 (31.1%) |
| Visual Fields | -- | 7/45 (15.6%) |
| Other | -- | 10/45 (22.2%) |
|
| ||
| Subjective Assessment | ||
|
| ||
| Tested | ||
| Yes | 75 (88.2%) | 68 (80.0%) |
| No | 10 (11.8%) | 17 (20.0%) |
|
| ||
| Method1 | ||
| Patient-Report | 67/75 (89.3%) | 61/68 (89.7%) |
| Provider-Report | 11/75 (14.7%) | 10/68 (14.7%) |
| Other | 10/75 (13.3%) | 8/68 (11.8%) |
|
| ||
| Accommodations | ||
|
| ||
| Offers Accommodations | ||
| Yes2 | 19 (22.3%) | 32 (37.6%) |
| No | 65 (76.5%) | 52 (61.2%) |
| Don’t Know | 1 (1.2%) | 1 (1.2%) |
Objective and subjective assessments are not mutually exclusive (i.e. studies could be counted for both)
Multiple methods may have been used by each study for objective or subjective assessment of sensory impairment
Hearing accommodations: amplification device (e.g. Pocket Talker); vision accommodations: reading aloud, magnification glass, enlarged type
Not all respondents who selected “other” reported the type of assessment method. Among those that did, objective methods included but were not limited to the following: “reports from community based subjects”, “distortion product otoacoustic emission”, “central auditory processing”, “dichotic digits”, “Canadian triple digit test”, “speech in noise test”, “Speech Audiometry”, “Masking Level Difference Test”, and “Synthetic Sentence Identification with ipsilateral competitive message” (hearing); “tests for accommodation”, “tonometry”, “retinal camera”, “Sussex vision test card”, “depth perception”, “OCT and Angio-OCT”, “Retinography with Autofluorescence”, or “Fissure Lamp Scan” (vision). Other specified subjective methods of assessment included but were not limited to “linking to medical health records”, “interviewer assessed”, or “proxy responses” for hearing; and “interviewer assessed”, “proxy responses”, “medical reports”, and “caregiver responses” for vision.
A majority of the cohorts (n=75; 88%) assessed both hearing and vision, either subjectively or objectively; a slightly smaller proportion of cohorts (n=66; 78%) assessed both hearing and vision subjectively. Less than one-third of the cohorts (n=27; 32%) assessed both hearing and vision objectively, while nearly half of the cohorts (n=34; 40%) assessed neither hearing nor vision objectively (Table 2).
Table 2.
Any, subjective and objective assessment of hearing and vision
| Vision | ||||||||
| Any | Subjective | Objective | ||||||
| Yes | No | Yes | No | Yes | No | |||
| Hearing | Any | Yes | 75 (88.2%) | 2 (2.4%) | -- | -- | -- | -- |
| No | 3 (3.5%) | 5 (5.9%) | -- | -- | -- | -- | ||
| Subjective | Yes | -- | -- | 66 (77.6%) | 9 (10.6%) | -- | -- | |
| No | -- | -- | 2 (2.4%) | 8 (9.4%) | -- | -- | ||
| Objective | Yes | -- | -- | -- | -- | 27 (31.8%) | 5 (5.9%) | |
| No | -- | -- | -- | -- | 18 (21.2%) | 34 (40.0%) | ||
Missing n=1
Study exclusion criteria
Table 3 outlines cohorts’ exclusion criteria for neurocognitive testing. A total of 6 (7.1%) cohorts had either hearing or vision exclusion criteria. Five out of 85 cohorts (5.9%) reported exclusion criteria based on hearing impairment: One cohort excluded participants with severe auditory “handicap” at the start of the study as well as participants who were mute or deaf at each test wave; two cohorts excluded participants who were deaf; one cohort excluded participants who demonstrated no hearing ability even while using an aid; and one cohort excluded participants who were unable to communicate sufficiently to be consented. Six (7.1%) cohorts reported exclusion criteria based on visual impairment. Five were the same cohorts that reported exclusion criteria based on hearing impairment: two cohorts excluded participants who were blind; one cohort excluded participants with severe visual impairment at the start of the study; one cohort excluded participants demonstrating no vision ability even while using an aid; and one cohort excluded participants who were unable to communicate sufficiently to be consented. One additional cohort excluded participants with corrected vision worse than 20/30 on the Snellen Eye Chart. Finally, 8 (9.4%) cohorts reported exclusion criteria based on cognition. Four (4.7%) cohorts reported on general exclusion criteria (e.g. age, language, area of residence, citizenship, community-dwelling status), some of which overlapped with exclusion criteria related to cognition (e.g. prior diagnosis of dementia).
Table 3.
Study exclusion criteria for hearing, vision and cognitive impairment
| Sensory1 | Cognition | General | |
|---|---|---|---|
| Hearing | Vision | ||
| • Severe auditory “handicap”, mute/speechless, or deaf • Deaf (n=2) • No hearing ability (even with aid) • Unable to communicate sufficiently to be consented |
• Severe visual “handicap” • Corrected vision worse than 20/30 on Snellen Eye Chart • Blind (n=2) • No vision ability (even with aid) • Unable to communicate sufficiently to be consented |
• MMSE < 26 • MMSE < 242 • MMSE < 10 • Dementia, suspected dementia, or “mental retardation” • Severe dementia • Does not understand (most of) the questions • Unable to hear the questions, not alert, or agitated |
• Institutionalized participants (includes those with severe dementia) • Diagnosed dementia, serious cardiovascular or cerebrovascular disease, or psychiatric conditions • “Too sick to participate” • 2+ domains of disability, institutionalized, or MMSE<24 |
MMSE = Mini Mental State Examination
Text in quotations are taken directly from survey responses; otherwise, paraphrased. Each bullet point refers to exclusion criteria listed in one cohort study, unless indicated otherwise (e.g. “n=2”).
Five studies had both hearing and vision exclusion criteria; 1 study only had vision exclusion criteria (N=6 with any exclusion criteria).
Participants with MMSE<24 would be administered a shorter version of the comprehensive protocol, and their cognitive data would be brought to consensus meeting for discussion.
Accommodations
Nineteen (22%) cohorts offered hearing accommodations and 32 (38%) cohorts offered vision accommodations during cognitive testing (Table 1). Only 17 (20%) cohorts offered accommodations for both hearing and vision, while more than half (n=50; 59%) offered accommodations for neither (Table 4).
Table 4.
Provision of hearing and vision accommodations
| Vision | |||
| Yes | No | ||
| Hearing | Yes | 17 (20.0%) | 2 (2.4%) |
| No | 15 (17.6%) | 50 (58.8%) | |
Missing n=1
DISCUSSION
Based on surveys completed for cohort studies collecting cognitive data among older adults, we observed variation in methods used to assess hearing and vision, exclusion criteria, and accommodation practices for sensory impairment during cognitive testing. Of the cohorts that measured sensory impairment, a majority used subjective methods such as self- and provider-report. A higher proportion of cohorts surveyed assessed vision impairment compared to the proportion that assessed hearing impairment, and a slightly higher proportion of cohorts offered vision accommodations compared to the proportion that offered hearing accommodations. These findings suggest a lack of standardization for assessing sensory impairment across cohort studies of cognitive function in older adults, and a call for a consensus by experts in the field is needed. To our knowledge, this is the first study to systematically survey cohort studies collecting neurocognitive data in older adults to describe protocol considerations for sensory impairment.
Neurocognitive tests from large, population-based studies are used to derive estimates of ADRD prevalence and identify at-risk groups. Valid estimates and projections of ADRD prevalence are needed for public health planning, etiologic research aimed at identifying approaches to reduce ADRD and cognitive decline risk, and the targeting of interventions to prevent or slow progression of these outcomes. However, completion of neurocognitive tests commonly used in these studies requires a baseline threshold of hearing and/or vision abilities. Given the high prevalence of sensory impairment in older adults and hypotheses of a potential causal effect of sensory impairment on cognition,9 the outright exclusion of participants with sensory impairment due to inability to access the test materials may lead to the exclusion of those at high risk of ADRD and consequently underestimate the prevalence of ADRD/cognitive impairment in those samples. Our study found that most cohorts did not exclude participants based on sensory impairment and that, encouragingly, more than two-thirds of the cohorts assessed both hearing and vision. This suggests that most cohorts are making an effort to assess sensory function and to include participants with sensory impairment where possible.
Nonetheless, standard methods for assessing cognitive function in research require the ability to hear or see test materials. If a participant usually wears hearing aids or glasses, some protocols explicitly document that they should be instructed to wear them during testing. However, glasses and hearing aids may be forgotten at home. Studies may consider reminding participants to bring them to their visit through a telephone call or letter sent in advance of the visit, or providing voice amplifiers and magnifying glasses on site. In our survey, severe sensory impairment (e.g., no hearing/vision ability even with aid) was an exclusion criterion for neurocognitive testing in 8% of the cohorts. Although not explicitly stated in the protocol for the remaining cohorts, it seems likely that, at the discretion of the psychometrist, participants with a similarly severe sensory impairment would also be excluded from testing because of an inability to access test materials. Yet exclusion results in an important inequity in the way in which cognitive function is assessed in research studies in older adults. This inequity may result in a cycle that perpetuates a lack of evidence on which to base policy and clinical decision-making. Although this phenomenon is well-described in other research areas, such as the exclusion of pregnant women from participation in randomized controlled trials,18 the impact of exclusion of older adults with disabilities from research on ADRD and cognitive function has heretofore received little attention.
As an alternative to exclusion, cohorts may consider other options to address sensory impairment in cognitive testing, such as the provision of sensory accommodations (e.g., large print testing materials, pocket-talkers) during testing. Our study found that 34 (40%) of the cohorts surveyed provided accommodations for at least hearing or vision, and that one-fifth provided accommodations for both hearing and vision impairment. The most common accommodations included an amplification device for hearing; reading aloud, magnification glass, and enlarged type for vision. If accommodations for sensory impairment fit within the overall goals of a particular study, supplementing neurocognitive testing with simple, standard assessments of hearing and vision, and provision of accommodations where indicated, would serve to promote inclusion and equity. Reasons for not providing accommodations (e.g., participant refusal) should also be documented and reported.
Alternatively, cohorts might consider providing accommodations to all participants, irrespective of the results of sensory testing. This universal approach could potentially help reduce any stigma participants may experience related to sensory impairment. It is also in keeping with the goal of many current protocols that explicitly strive to create an optimal testing environment for the participants. For example, many protocols specify administration of the cognitive tests face-to-face in a quiet room (e.g., limited noise from nearby sources such as conversations, ventilation systems, etc.) and in multiple sensory modalities (e.g. in written and verbal forms) where possible. Systematic documentation of these procedures across all studies would provide a more comprehensive picture of how sensory impairment is handled in cognitive testing. However, it should be acknowledged that study authors’ ability to document these methods in manuscripts is likely limited by word count constraints in peer-reviewed publications. To facilitate a clearer understanding among researchers of whether and how participants were excluded from studies, journals may consider providing supplement options specifically for procedures pertaining to equitable inclusion of study participants.
It should also be noted that the decision of whether to provide accommodations should reflect the specific goals of the study. Sensory impairment is hypothesized to causally contribute to ADRD and cognitive decline through several non-mutually exclusive pathways. One such pathway is through increased cognitive load, which is the additional cognitive processing effort due to sensory impairment.19 For example, hearing requires successful transduction and encoding of sound in the cochlea, followed by central processing of that auditory signal in the brain. With hearing impairment, encoding may be impaired, resulting in a degraded auditory signal. Therefore, successful decoding of this degraded signal may be at the expense of resources that might have otherwise been used for encoding speech content into memory. If the goal of the study is to evaluate the association between hearing impairment and worse cognitive performance, provision of accommodations may be suboptimal since accommodations may directly mitigate this pathway. Where possible, studies may consider conducting tests with and without accommodations in order to assess the contribution of sensory impairment.
In addition to accommodations, other options may provide alternatives to exclusion from cognitive testing. One cohort in this analysis skipped some tests when the participant had severe hearing or vision impairment instead of excluding participants with any sensory impairment. However, it should be acknowledged that changing any test in a battery fundamentally changes the construct being measured, and as such, instruments should ideally be standardized across all participants. Another alternative is to develop normative cognitive data for older adults with sensory impairment, though this would not provide any information on whether differences in performance are due to biological correlates of ADRD associated with sensory impairment or due to the participant’s diminished inability to read or hear the test questions. Neumann and colleagues (2017) suggested using nonvisual and nonverbal cognitive testing to narrow the gap of missing data related to older adults with significant sensory or motor impairment, but it is unclear how this could be implemented on a wide scale. The use of proxy instruments such as the Eight-item Informant Interview to Differentiate Aging and Dementia (AD8) may be used for studies specific to ADRD.20 However, these instruments do not allow for the detection of milder forms of cognitive impairment.
In general, administration of cognitive tests on a computer may allow for larger font sizes as well as higher amplification with a neutral voice. For studies that want to ensure that participants understand spoken speech, one option could be the procedure employed by both the Aging and Cognitive Health Evaluation in Elders (ACHIEVE) randomized study and the Atherosclerosis Risk in Communities Study (ARIC).21,22 Prior to testing, the examiner reads five sentences (three key words in each) at a volume and pace commensurate with cognitive testing. If participants are unable to correctly repeat back 13 or more key words, a second set of sentences are read at a louder volume. If the participant correctly repeats back 13 or more key words, cognitive testing proceeds at the louder volume. If the participant is again unable to repeat back 13 or more of the 15 keywords, testing proceeds at the louder volume, supplemented by written instructions. On balance, there are several possible solutions to this problem with a combination of benefits and drawbacks, but simply excluding participants based on sensory impairment could perpetuate inequities and potentially lead to missed opportunities for intervention.
In our survey, we assessed measurement of sensory function both through objective measures and self-report. While objective assessments such as audiometry and visual acuity can identify the presence or level of biological impairment, self-report takes into account the participants’ perception of how the impairment affects their everyday function, a construct more akin to disability.23 Although correlated, the two measures are not identical; self-report and objective measures differ in their ability to identify individuals with and without sensory impairment.24 Overall, prior studies suggest that self-report may identify ~70% of individuals with measurable audiometric loss24, and similarly, multiple vision tests (better visual acuity, contrast sensitivity, stereoacuity, and visual fields) are significantly associated with self-rated vision status.25 Importantly, demographic factors strongly related to cognition may influence how a person self-reports. Comparing self-report to audiometry, older age is associated with underestimation of hearing impairment, but higher education is associated with overestimation.24 Hence the choice of assessment type should be within the context of each study’s goals and which construct they wish to measure. Self-reported questions are relatively quick and easy to administer but may not identify everyone with impairment and may also identify some individuals as having hearing impairment, when their audiometric thresholds would fall within the normal hearing range. Our findings emphasize the need for systematically documenting and reporting the methods used to assess sensory impairment in studies of cognitive function, including the rationale for using said methods.
A major limitation of our study is the low survey response rate (<50%). Cohorts that are interested in research questions related to sensory impairment and those that prioritized assessment of sensory impairments during protocol development may have been more inclined to respond to the survey. If this response bias exists, it may lead to an overestimation of the proportion of cohorts that assessed sensory impairment. However, our primary goal for this study was to survey what is currently being done in the field. Our data show that there is a lack of standardization across protocols, which would likely not change with an expanded sample of cohorts. Another limitation was that we had surveyed cohorts on whether they offered accommodations, but it was unclear what proportion of participants in each cohort needed those accommodations. If a cohort enrolled a healthy subgroup of older adults, they may not have offered accommodations because there was not a need to do so. Additionally, the age range for those who are considered “older adults” may vary between cohorts. Since sensory impairment is more prevalent among those who are older, this variation in definition could affect whether cohorts assessed sensory impairment and whether they offered accommodations. Additionally, our review was restricted to prospective studies, and so well-conducted cross-sectional studies (e.g., NHANES) were not included. Finally, we acknowledge that the actual methods used in studies may vary from the protocol. Survey respondents had an option to request a phone call to discuss protocol deviations, but we cannot assure all protocol deviations were recorded. Relatedly, while we can deduce from study protocols that some participants were excluded from cognitive data collection due to sensory loss, it is unclear how many participants were actually impacted in each study. Future studies should evaluate full study protocols when available. Despite these limitations, our study is novel and provides insight into current practices for collection of cognitive data among older adults with sensory impairment.
In summary, our findings call for the development of standardized methods to collect cognitive data and the need to identify best practices to harmonize protocols across cohorts. Studies of cognitive function in older adults should carefully consider sensory impairment, systematically document how sensory impairment was assessed, and report these methods in peer-reviewed literature. Objective and subjective assessments of sensory impairment measure different constructs, which should be taken into account when deciding which one to use in combination with cognitive testing. Studies should also aim to report whether and which accommodations were used as well as how it was determined that accommodations were needed. Future work and consensus-building are needed to determine and develop better methods to assess cognitive function in older adults with sensory impairment.
Supplementary Material
RESEARCH IN CONTEXT.
Systematic review:
Sensory loss is independently associated with accelerated cognitive decline and increased risk of dementia. However, it is unclear whether these associations are due to causal mechanistic pathways such as neurodegeneration and social isolation or due to bias related to cognitive testing. There have been no formal analyses of research protocols to identify the degree of variation in methods used to collect and analyze cognitive testing data in research settings among older adults with sensory impairments.
Interpretation:
Our findings showed significant variation in exclusion criteria related to hearing or vision impairment, methods of assessment, and accommodation practices.
Future directions:
Our findings call for the development of standardized methods to collect cognitive data and the need to identify best practices to harmonize protocols across cohorts. Studies should carefully and systematically document how sensory impairment was assessed, whether and which accommodations were provided, and report these methods in peer-reviewed literature.
Acknowledgements
This work was supported by National Institute on Aging (NIA) grants R21AG060243 (BKS and JAD), K01AG052640 (BKS) and K01AG054693 (JAD), and the James E. West Fellowship from the Acoustical Society of America (DV). The authors thank participating PIs and study coordinators.
Role of the funding source
The funders of this study had no role in study design, data collection, data analysis, data interpretation, or writing of the report. The corresponding author had full access to all data in the study and had final responsibility for the decision to submit for publication.
Declarations of interest: L.R. received a grant from the National Eye Institute / National Institutes of Health (UG1-EY020522-08) and consulting fees from Sciome. D.V. received the James E West Fellowship from the Acoustical Society of America. M.C.C. received grants from the National Institute on Aging (R01AG066153 and R01AG055404) and the Department of Defense (W81XWH1910730); and honoraria from Cal State North Ridge, the International Neuropsychological Society (INS), and the AARP Staying Sharps scientific advisory board. B.K.S received grants from the National Institutes of Health / National Institute on Aging (R21AG060243 and K01AG052640) and consulting fees from the American Foundation for the Blind. J.A.D. received grants from the National Institute of Health / National Institute on Aging (R21AG060243 and K01AG054693) and honoraria from speechpathology.com, audiologyonline.com, and Rice University.
REFERENCES
- 1.Lin FR, Niparko JK, Ferrucci L. Hearing loss prevalence in the United States. Arch Intern Med. 2011. Nov 14;171(20):1851–2. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Swenor BK, Ramulu PY, Willis JR, Friedman D, Lin FR. The prevalence of concurrent hearing and vision impairment in the United States. JAMA Intern Med. 2013. Feb 25;173(4):312–3. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Livingston G, Huntley J, Sommerlad A, Ames D, Ballard C, Banerjee S, et al. Dementia prevention, intervention, and care: 2020 report of the Lancet Commission. Lancet. 2020. Aug 8;396(10248):413–46. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Loughrey DG, Kelly ME, Kelley GA, Brennan S, Lawlor BA. Association of Age-Related Hearing Loss With Cognitive Function, Cognitive Impairment, and Dementia: A Systematic Review and Meta-analysis. JAMA Otolaryngol Head Neck Surg. 2018. Feb 1;144(2):115–26. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Chen SP, Bhattacharya J, Pershing S. Association of Vision Loss With Cognition in Older Adults. JAMA Ophthalmol. 2017. Sep 1;135(9):963–70. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Zheng DD, Swenor BK, Christ SL, West SK, Lam BL, Lee DJ. Longitudinal Associations Between Visual Impairment and Cognitive Functioning: The Salisbury Eye Evaluation Study. JAMA Ophthalmol. 2018. Sep 1;136(9):989–95. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Neumann D, Robinski M, Mau W, Girndt M. Cognitive Testing in Patients with CKD: The Problem of Missing Cases. Clin J Am Soc Nephrol. 2017. Mar 7;12(3):391–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Deal JA, Gross AL, Sharrett AR, Abraham AG, Coresh J, Carlson M, Griswold M, Mosley T, Power MC, Ramulu P, Reed NS, Lin FR, Swenor BK. Hearing impairment and missing cognitive test scores in a population-based study of older adults: The Atherosclerosis Risk in Communities neurocognitive study. Alzheimers Dement. 2021. Apr 12. doi: 10.1002/alz.12339. Epub ahead of print. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Whitson HE, Cronin-Golomb A, Cruickshanks KJ, Gilmore GC, Owsley C, Peelle JE, et al. American Geriatrics Society and National Institute on Aging Bench-to-Bedside Conference: Sensory Impairment and Cognitive Decline in Older Adults. J Am Geriatr Soc. 2018. Nov;66(11):2052–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Pirozzo S, Papinczak T, Glasziou P. Whispered voice test for screening for hearing impairment in adults and children: systematic review. BMJ. 2003. Oct 25;327(7421):967. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Torres-Russotto D, Landau WM, Harding GW, Bohne BA, Sun K, Sinatra PM. Calibrated finger rub auditory screening test (CALFRAST). Neurology. 2009. May 5;72(18):1595–600. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Walker JJ, Cleveland LM, Davis JL, Seales JS. Audiometry screening and interpretation. Am Fam Physician. 2013. Jan 1;87(1):41–7. [PubMed] [Google Scholar]
- 13.Bailey IL, Lovie JE. New design principles for visual acuity letter charts. Am J Optom Physiol Opt. 1976. Nov;53(11):740–5. [DOI] [PubMed] [Google Scholar]
- 14.Lovie-Kitchin JE. Validity and reliability of visual acuity measurements. Ophthalmic Physiol Opt. 1988;8(4):363–70. [DOI] [PubMed] [Google Scholar]
- 15.Legge GE, Ross JA, Luebker A, LaMay JM. Psychophysics of reading. VIII. The Minnesota Low-Vision Reading Test. Optom Vis Sci. 1989. Dec;66(12):843–53. [DOI] [PubMed] [Google Scholar]
- 16.Pelli DG, Robson JG, Wilkins AJ. The design of a new letter chart for measuring contrast sensitivity. Clinical Vision Sciences. 1988;2(3):187–199. [Google Scholar]
- 17.Gaasterland DE, Ederer F, Sullivan EK, Caprioli J. Advanced Glaucoma Intervention Study. 2. Visual field test scoring and reliability. Ophthalmology. 1994. Aug;101(8):1445–55. [PubMed] [Google Scholar]
- 18.Beigi RH, Krubiner C, Jamieson DJ, Lyerly AD, Hughes B, Riley L, et al. The need for inclusion of pregnant women in COVID-19 vaccine trials. Vaccine. 2021. Feb 5;39(6):868–70. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Tun PA, McCoy S, Wingfield A. Aging, hearing acuity, and the attentional costs of effortful listening. Psychol Aging. 2009. Sep;24(3):761–6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Galvin JE, Roe CM, Powlishta KK, Coats MA, Muich SJ, Grant E, et al. The AD8: a brief informant interview to detect dementia. Neurology. 2005. Aug 23;65(4):559–64. [DOI] [PubMed] [Google Scholar]
- 21.Deal JA, Goman AM, Albert MS, Arnold ML, Burgard S, Chisolm T, et al. Hearing treatment for reducing cognitive decline: Design and methods of the Aging and Cognitive Health Evaluation in Elders randomized controlled trial. Alzheimers Dement (N Y). 2018;4:499–507. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.The Atherosclerosis Risk in Communities (ARIC) Study: design and objectives. The ARIC investigators. Am J Epidemiol. 1989. Apr;129(4):687–702. [PubMed] [Google Scholar]
- 23.Verbrugge LM, Jette AM. The disablement process. Soc Sci Med. 1994. Jan;38(1):1–14. [DOI] [PubMed] [Google Scholar]
- 24.Kamil RJ, Genther DJ, Lin FR. Factors associated with the accuracy of subjective assessments of hearing impairment. Ear Hear. 2015. Jan;36(1):164–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.El-Gasim M, Munoz B, West SK, Scott AW. Associations between self-rated vision score, vision tests, and self-reported visual function in the Salisbury Eye Evaluation Study. Invest Ophthalmol Vis Sci. 2013;54(9):6439–6445. Published 2013 Sep 27. doi: 10.1167/iovs.12-11461 [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.