HI-MoCA: Cognitive Test for Hearing Impaired Adults

The Laryngoscope
C 2017 The American Laryngological,
V
Rhinological and Otological Society, Inc.
TRIOLOGICAL SOCIETY
CANDIDATE THESIS
Development of Cognitive Screening Test for the Severely
Hearing Impaired: Hearing-Impaired MoCA
Vincent Y. W. Lin, MD, FRCSC; Janet Chung, MD, FRCSC; Brandy L. Callahan, PhD, Psych;
Leah Smith, MA, CCRA; Nils Gritters, BSc (Hons); Joseph M. Chen, MD, FRCSC;
Sandra E. Black, OC, O.Ont. MD, FRSC, FAAN, FANA; Mario Masellis, MSc, MD, PhD, FRCPC
Objectives: To develop a version of the Montreal Cognitive Assessment (MoCA) to be administered to the severely hearing
impaired (HI-MoCA), and to assess its performance in two groups of cognitively intact adults over the age of 60.
Study Type: Test development followed by prospective subject recruitment.
Methods: The MoCA was converted into a timed PowerPoint (Microsoft Corp., Redmond, WA) presentation, and verbal
instructions were converted into visual instructions. Two groups of subjects over the age of 60 were recruited. All subjects passed
screening questionnaires to eliminate those with undiagnosed mild cognitive impairment. The first group had normal hearing
(group 1). The second group was severely hearing impaired (group 2). Group 1 received either the MoCA or HI-MoCA test (T1).
Six months later (T2), subjects were administered the test (MoCA or HI-MoCA) they had not received previously to determine
equivalency. Group 2 received the HI-MoCA at T1 and again at T2 to determine test–retest reliability.
Results: One hundred and three subjects were recruited into group 1, with a score of 26.66 (HI-MoCA) versus 27.14
(MoCA). This was significant (P < 0.05), but scoring uses whole numerals and the 0.48 difference was found not clinically
significant using post hoc sensitivity analyses. Forty-nine subjects were recruited into group 2. They scored 26.18 and 26.49
(HI-MoCA at T1 and T2). No significance was noted (P > 0.05), with a test–retest coefficient of 0.66.
Conclusion: The HI-MoCA is easy to administer and reliable for screening cognitive impairment in the severely hearing
impaired. No conversion factor is required in our prospectively tested cohort of cognitively intact subjects.
Key Words: hearing loss, dementia, screening, elderly.
Level of Evidence: 1b.
Laryngoscope, 127:S4–S11, 2017
INTRODUCTION
The association between hearing loss and dementia
was first proposed in 1989 in a case-control study published by Uhlmann et al. in the Journal of the American
Medical Association.1 In this article, 100 patients with
Alzheimer’s-type dementia were matched to 100 control
subjects without dementia with respect to age, sex, and
education. They found that hearing loss was significantly and independently correlated with the severity of
cognitive dysfunction. This important relationship was
From the Department of Otolaryngology–Head & Neck Surgery
(V.Y.W.L., JANET C., L.S., N.G., JOSEPH C.); the Division of Neurology (B.L.C.,
S.E.B., M.M.); the Sunnybrook Research Institute (V.Y.W.L., B.L.C., L.S., N.G.,
JOSEPH C., S.E.B., M.M.), Sunnybrook Health Sciences Centre; and the
Department of Otolaryngology–Head & Neck Surgery, Faculty of
Medicine, University of Toronto (V.Y.W.L., JANET C., JOSEPH C.), Toronto,
Ontario, Canada.
Editor’s Note: This Manuscript was accepted for publication
February 21, 2017.
This study was partially funded by MED-EL, GmbH via a research
grant to the primary author (V.Y.W.L.). The authors have no other funding, financial relationships, or conflicts of interest to disclose.
Send correspondence to Vincent Y.W. Lin, MD, FRCSC, Associate
Professor, Associate Scientist, Department of Otolaryngology–Head &
Neck Surgery, Sunnybrook Research Institute, Sunnybrook Health
Sciences Centre, Institute of Medical Sciences, Faculty of Medicine,
University of Toronto, M1-102, 2075 Bayview Avenue, Toronto, Ontario,
M4N 3M5. E-mail: [email protected]
sparsely studied until 2011, when Lin et al. released a
series of groundbreaking publications. Their work established that age-related hearing loss is independently associated with worse cognitive function2 and incidental
dementia.3 They also established that the severity of hearing loss correlates with a higher risk of developing dementia; those with mild, moderate, and severe hearing loss,
respectively, had a 1.9-, 3.0-, and 4.9-fold increased risk of
developing dementia. Although an etiological link has yet
to be determined, an important question remains: how do
we best identify hearing-impaired patients who are at risk
of developing, or who already have, early dementia? A
screening measure, appropriate for the hearing-impaired
population, would allow for earlier environmental and
pharmacological intervention in individuals with early
cognitive impairment/dementia. It also would facilitate
identification of such patients for studies attempting to
unravel this etiological link.
Prevalence and Cost of Dementia
DOI: 10.1002/lary.26590
In Canada, the prevalence of dementia is rising due
to the rapidly growing aging segment of the population.
Wong et al. used information from the 2010/2011 Canadian Community Health Survey, the 2011/2012 Survey of
Neurological Conditions in Institutions in Canada, and
the 2011 Survey on Living with Neurological Conditions
in Canada and calculated that for individuals over the age
Laryngoscope 127: 2017
Lin et al.: Development of Hearing-Impaired MoCA
S4
of 45, 0.8% of those living in private households and 45%
of patients in long-term residential care facilities had a
diagnosis of dementia.4 There was an estimated 53,100
patients living in long-term residential care with dementia, and the prevalence rose with increasing age. The
impending socioeconomic costs of dementia have been
long recognized. In an article published in 1994, Ostbye
et al. calculated the net cost of dementia in Canada to be
3.9 billion dollars.5 Those costs included the cost of care in
long-term institutions and medical/social assistance. More
recently in 2011, the Alzheimer’s Society of Canada calculated the cost in 2011 to be $33 billion a year. This cost was
estimated to rise to $293 billion a year by 2040.
Caregiver Burden
As the severity of dementia in an individual advances, there is an increasing reliance on caregivers for maintaining their activities of daily living. This usually falls
upon immediate family members, but as dementia worsens and other complications (e.g., immobility and behavioral disturbances) emerge, patients are often transferred
to long-term care facilities. In Canada, family caregivers
in up to 90% of these situations provide in-home care for
people with dementia.6 In the United States, there are
over 15 million unpaid caregivers who collectively provide
17.9 billion hours of assistance to the estimated 5.3 million
patients with Alzheimer’s disease (AD).7 There undoubtedly is a large amount of burden on the caregivers of
dementia patients both personally and economically.8,9 It
appears that caregiver burden is not only a first-world
issue, Pattanayak et al. also describe a similar increasing
concern regarding caregiver burden in India.10
Diagnosis of Dementia
Although the prevalence of dementia is rising, its diagnosis still can remain challenging, especially in individuals
with primary sensory deficits. A diagnosis of dementia, also
called major neurocognitive disorder, requires subjective and
objective evidence of cognitive decline sufficiently marked to
impact activities of daily living.11 In patients with a high likelihood of dementia due to AD, this decline must affect at least
two of the following domains: 1) memory, 2) communication/
language, 3) attentiveness/focus on a program or subject, 4)
reasoning/judgment, and 5) visual perception.12
A recent systematic review performed by Ngo et al.
reviewed dementia practice guidelines from 12 working
groups from a variety of countries, including Canada, the
United States, the United Kingdom, Korea, Australia,
Singapore, Malaysia, and European countries.13 General
consensus was found in four main areas: 1) cognitive testing, 2) comorbid conditions, 3) clinical diagnosis, and 4)
imaging/specialized tests. Some highlights include testing
cognition using only standardized screening tools, such as
the Montreal Cognitive Assessment (MoCA); running a
standard battery of bloodwork, including complete blood
count, thyroid-stimulating hormone, electrolytes, calcium,
fasting blood glucose, vitamin B12, and folate; making
diagnoses of dementia using internationally agreed upon
consensus criteria; and using at least one imaging
Laryngoscope 127: 2017
modality. Although consensus was seen in a number of
guideline recommendations, diverging recommendations
were still evident in their review. Clearly the diagnosis of
dementia and its various subtypes requires the skill of
experienced clinicians relying on a wide breadth of standardized tests to aid them.
Diagnosis of Mild Cognitive Impairment
Substantial effort has been made in recent years to
move toward earlier diagnosis and disease management.
Mild cognitive impairment (MCI), also called mild neurocognitive disorder by Diagnostic Statistical Manual of
Mental Disorders-V criteria,11 describes cognitive decline
that is beyond expected age-related change and is
assumed to reflect incipient dementia, although not severe
enough to interfere with independence in functional abilities.14 Core clinical criteria include: 1) concern regarding
a change in cognition, 2) impairment in one or more cognitive domains, 3) preservation of independence in functional abilities, and 4) not demented with no evidence of
significant impairment in social or occupational functioning.14 MCI is of particular interest to clinicians not only
because these patients can still function at a fairly high
level but because it is felt that interventions may be most
effective in these early-stage patients.15
According to the Alzheimer’s Society of Canada:
Family, friends and the person with MCI may notice
these changes and they can be objectively measured in
cognitive tests, but the changes in cognitive abilities
are usually not serious enough to interfere with daily
life and independence. A person with MCI has an
increased risk of developing AD or another dementia,
but some people with MCI remain stable and others
show improved cognitive abilities over time. There is no
single cause or outcome for people diagnosed with MCI.
A wide range of cognitive abilities can be impaired, but
for the majority of people who have MCI memory is the
cognitive ability most affected.
Superiority of MoCA
The gold standard to identify cognitive changes that
may be indicative of dementia remains formal neuropsychological assessment involving the standardized administration of a broad battery of measures by a trained
professional.16 To identify individuals who may benefit
from such in-depth neuropsychological testing, panels of
experts worldwide have issued formal recommendations
for general practitioners to screen for dementia in highrisk groups.17 Historically, the Mini-Mental State Examination (MMSE) has been widely used for cognitive screening. It is brief and easy to administer and score. Generally,
a score of 26 or greater out of 30 is considered normal, taking into account education level. However, multiple studies have demonstrated the MMSE’s low sensitivity in the
screening of cognitive impairment, especially in those with
mild cognitive impairment MCI.17–19
Indeed, the vast majority of patients with a neuropsychologically confirmed clinical diagnosis of MCI score
Lin et al.: Development of Hearing-Impaired MoCA
S5
above the normal cutoff of 26 on the MMSE.20 The MoCA
was developed as a more sensitive screening tool for
patients with MCI and also is useful for more severe
stages of dementia. The MoCA also is easy to administer
and score, with normal subjects scoring 26 or greater out
of a total score of 30 (www.mocatest.org).20 It consists of 11
distinct test components measuring executive functions,
language, attention, abstraction, verbal fluency, calculation, concentration, visuospatial skills, orientation, and
delayed recall. Validation studies involved normal controls, patients with confirmed MCI, and those with confirmed AD.19,20 Patients in all groups were tested with
both the MMSE and MoCA, and the results indicated that
the MoCA’s sensitivity in detecting MCI was 90%, far
superior to the MMSE’s sensitivity (18%). The authors
attributed the MoCA’s significant advantage to its testing
of executive functions, higher-level language abilities, and
complex visuospatial processing. It also has been used
effectively to identify mild cognitive impairment and
dementia in patients with Parkinson’s disease.21
Need for HI-MoCA
It has been demonstrated that healthy community
living adults (mean age 71) with hearing impairment
perform more poorly on the MoCA, even if the hearing
impairment is not severe enough to prohibit standard
verbal administration of the test.22 Severe hearing
impairment clearly prevents most forms of standardized
neuropsychological test administration from being properly and accurately administered. Thus, test performances are negatively impacted.23
For patients with severe hearing impairment, the administration of the MoCA as a sensitive screening test for MCI and
dementia is currently impossible given that the test requires
verbal instructions. In our adult cochlear implantation program, there has been a substantial increase in the number of
older patients (> 55 years old) who are being considered for
cochlear implantation, primarily due to the expansion of
implantation criteria and the well-documented aging of our
population. Informally during consultation by either the
cochlear implant audiologist or surgeon, the patient’s family
may raise concerns regarding the possibility of cognitive
impairment, or it may be noted by the clinician during the
encounter. In these circumstances, no validated and recognized
screening test for cognitive impairment exists that could be
used to assess severely hearing-impaired individuals.
Furthermore, it is critical to identify if a patient
suffers from MCI or dementia because outcomes after
surgery often are worse for this subgroup of elderly individuals from the perspective of developing delirium and
postoperative complications.24
Therefore, out of this clinical need, we sought to
adapt the MoCA for hearing-impaired patients (HI-MoCA)
and validate this modified test first in our group of cognitively intact subjects.
Study Objectives
Our study objectives were to develop a version of
the MoCA that could be administered to the severe
Laryngoscope 127: 2017
S6
hearing impaired, namely the hearing-impaired MoCA
(HI-MoCA), and to validate it in two groups of cognitively intact adults over the age of 60.
Study Aims
1. Develop a hearing-impaired version of the MoCA: the HI-MoCA.
a. Convert the verbal instructions into visual instructions
displayed on an interactive Microsoft Powerpoint (Microsoft Corp., Redmond, WA) presentation that is as internally consistent as possible with the original MoCA.
b. Have the HI-MoCA administered without any verbal cues
so the tester is present simply to ensure that the proper
visual directions are being followed.
2. Validate the HI-MoCA in two groups of cognitively intact
individuals. Cognition would be assessed with a validated
battery of cognitive screening tests.
a. Group 1: Normal hearing, cognitively intact
b. Group 2: Severe hearing loss, cognitively intact
Study Hypothesis
Our hypothesis is that our HI-MoCA will be clinically equivalent to the original MoCA in our cognitively
intact patients (with or without severe hearing loss).
MATERIALS AND METHODS
Study Methodology
HI-MoCA Test Development and Procedure. Permission
was obtained from Dr. Ziad Nasreddine, MD, to modify the
MoCA into an hearing-impaired MoCA (HI-MoCA). We also
obtained institutional ethics approval to complete this study.
To administer the HI-MoCA to subjects without any verbal
instructions or prompts, we converted the original verbal instructions into a visual PowerPoint (Microsoft Corp.) presentation on a
computer screen. The patients read the instructions, and after
acknowledging their understanding of the instructions, press a
button to advance the presentation. They are also given a series
of blank answer sheets to be completed as instructed by the
PowerPoint presentation. Once advanced, the instructions disappear and are replaced with instructions to complete the task on
the sheet of paper in front of them. Once the task is completed,
the completed answer sheet is corrected so answers cannot be
changed after the task is completed. The next blank answer sheet
then awaits the subject to complete while, the next task is
explained visually on the computer screen. These visual instructions are sufficient for sections on the MoCA, such as the visuospatial/executive tests (trails, copy cube, and clock draw) (Fig. 1).
Visual instructions also are used in the naming section (Fig. 2).
In the memory recall section, a series of words are presented at a rate of 3 seconds per word (face, velvet, church, daisy, red), comparable to the standard verbal administration of
the original MoCA. A blank answer sheet is used by the subject
to recall these words in writing. Immediately following this, the
completed answer sheet is taking away and subjects are directed to the next blank answer sheet for a second trial recall.
For the attention task requiring the participant to identify
the letter A among a string of letters by tapping the table, the
HI-MoCA visually presents a series of letters at a rate of 2
seconds between letters for the subject to observe and respond
to by tapping the table as well.
Finally, in the abstraction subtest, the HI-MoCA test
administrator has additional slides that can be used if required.
For example, the subject is asked to write down the similarity
between a banana and orange. The correct answer is “they are
Lin et al.: Development of Hearing-Impaired MoCA
Fig. 1. Screen capture of the visually described trails task for the Hearing-Impaired Montreal Cognitive Assessment, along with the corresponding patient worksheet [Color figure can be viewed in the online issue, which is available at www.laryngoscope.com.]
both fruit.” However, if the subject instead writes that both can
be eaten, the test administrator can advance to a screen that
displays the statement: “The answer we were looking for is that
an orange and a banana are both fruit,” similarly to the verbal
cues provided in the original MoCA.
Participants
Using the original MoCA validation study20 as a guideline
for sample sizes, 103 participants were prospectively recruited
from our adult cochlear implant program. Group 1 consisted of
103 subjects with normal hearing and normal cognition. Group
2 consisted of 49 subjects with severe hearing loss and normal
cognition.
Participants were all over the age of 60 and were excluded
if they had a family history of dementia or had a history of any
neurological disease. A detailed personal health history was
taken prior to enrollment in our study to ensure that patients,
both normal hearing and hearing impaired, were as cognitively
intact as possible. In addition, three screening questionnaires
developed for identifying cognitive impairment were administered, and patients also were excluded if they failed any of
these assessments. These three questionnaires were:
1. Personal Health Questionnaire, including hearing screening,
Developed in-house at Sunnybrook Health Sciences Centre
2. Activities of Daily Living Questionnaire by proxy (completed
by a spouse, partner, or caregiver)25
Fig. 2. Screen capture of the visually described animal-naming task for the Hearing-Impaired Montreal Cognitive Assessment, along with
the corresponding patient worksheet [Color figure can be viewed in the online issue, which is available at www.laryngoscope.com.]
Laryngoscope 127: 2017
Lin et al.: Development of Hearing-Impaired MoCA
S7
Fig. 3. Testing timeline for groups 1 and 2.
MoCA 5 Montreal Cognitive Assessment; HI-MoCA 5 Hearing-Impaired Montreal Cognitive Assessment. [Color figure can be viewed in the
online issue, which is available at www.laryngoscope.com.]
3. Alzheimer’s Society of Canada screening questionnaire for
primary care physicians26
Based on prior clinical experiences, our intake screening,
along with the three administered questionnaires, are effective
in excluding a significant majority of patients with early cognitive impairment who previously have not been diagnosed.
Subjects within group 1 (normal hearing, normal cognition)
were randomized to receive either the standard MoCA (n 5 43) or
HI-MoCA (n 5 60) during their first visit (T1). Approximately 6 to
8 months later (T2), subjects were retested with either the MoCA
or HI-MoCA, which they did not receive during their first visit.
This inter-test interval was selected to minimize any potential
learning effects from repeating the MoCA (Fig. 3).
Subjects in group 2 (severe hearing loss, normal cognition)
were given the HI-MoCA test during their first visit and subsequent visit 6 to 8 months later.
In total, eight versions of the HI-MoCA were evaluated.
In the earlier versions, a small pilot group of different subjects
(5–10) underwent testing and then interim analyses were performed, along with extensive discussions, among our research
group. Next, the HI-MoCA test was modified to incorporate
these changes. Many of the changes were made to help streamline the test administration and reduce the need for prompting
or tester intervention. This final data set was collected using
the eighth version of the HI-MoCA on subjects who were not
involved in the prior pilot studies.
Statistical Analyses
Descriptive statistics, including averages and standard
deviations (SD), were used to summarize the results of the
respondents, their MoCA/HI-MoCA scores, and the time between
tests in weeks.
Prior to selecting the type of statistical test used, we tested for normality. Because our data was normally distributed, a
paired t test was used to compare MoCA and HI-MoCA scores,
as well as HI-MoCA scores between groups. Statistical
significance was defined as a P < 0.05. When multiple comparisons were made in the MoCA subset analysis, the P value was
adjusted to account for these multiple comparisons (0.05 divided
by number of comparisons made).
A two-way repeated measures analysis of variance (ANOVA)
was run to determine whether the order of the testing—original
MoCA (T1), then HI-MoCA (T2 vs. HI-MoCA [T1]), and then
original MoCA (T2)—would impact the scoring for group
1. Analysis of the studentized residuals and the ShapiroWilk test were used to determine normality. Sphericity of the
interaction term (order and time) was assessed by Mauchly’s test,
and a priori it was determined that simple main effects would be
run if the interaction term was statistically significant.
Test–retest reliability was calculated using the standard
coefficient when testing between time periods in group 2 for the
HI-MoCA.
RESULTS
Results: Group 1 (Normal Hearing, Normal
Cognition) and Group 2 (Severe Hearing Loss,
Normal Cognition)
There were 103 subjects in group 1 who passed the
intake form and all three screening questionnaires, and
who completed both the original MoCA and HI-MoCA.
Of these, 43 subjects underwent the original MoCA, first
followed by the HI-MoCA 6 months later; and 60 subjects underwent the HI-MoCA, first followed by the original MoCA 6 months later (Table I). There was not
statistical difference between group 1 or group 2 with
regard to age and years of education.
In group 1, the average HI-MoCA score was 26.66
(range: 20–30, SD 5 2.542). The original MoCA score
was 27.14 (range: 21–30, SD 5 2.005). The original
MoCA scores were statistically significantly better than
the HI-MoCA scores, t(102) 5 22.11, P 5 0.037. This
TABLE I.
Demographic Data.
Male
Female
Years of Education
Average Age
Number of Weeks Between Tests
Group 1
(n 5 103)
29
74
15.65 (SD 5 2.25)
68.41 (SD 5 6.17)
28.93 (SD 5 6.17)
Group 2
23
26
15.04 (SD 5 2.40)
70.23 (SD 5 6.74)
29.11 (SD 5 5.97)
(n 5 49)
Laryngoscope 127: 2017
S8
Lin et al.: Development of Hearing-Impaired MoCA
TABLE II.
MoCA and HI-MoCA Subtest Scores.
Original MoCA
HI-MoCA
P Value
(sig < 0.007)
Visuospatial/executive
4.69 (SD 5 0.58)
4.54 (SD 5 0.73)
0.054
Naming
Memory
2.85 (SD 5 0.38)
3.74 (SD 5 1.15)
2.92 (SD 5 0.27)
4.09 (SD 5 1.14)
0.145
0.005
Group 1: Subtest
Attention
5.77 (SD 5 0.51)
5.64 (SD 5 0.79)
0.144
Language
Abstraction
2.42 (SD 5 0.69)
1.72 (SD 5 0.51)
2.08 (SD 5 0.87)
1.54 (SD 5 0.72)
0.001
0.026
Orientation
5.92 (SD 5 0.27)
5.82 (SD 5 0.49)
0.063
Significant results are bolded.
MoCA 5 Montreal Cognitive Assessment; HI-MoCA 5 Hearing-Impaired Montreal Cognitive Assessment; SD 5 standard deviation.
data suggests that the HI-MoCA is more challenging
because subjects scored on average 0.48 points lower
when compared to the MoCA. Exploratory follow-up
analyses were conducted to determine whether this
difference could be attributed to either one or several of
the subtests (Table II).
The 0.48-point difference between the HI-MoCA
and MoCA scores was not thought to be clinically significant given that the scoring of both tests is based on
whole numbers with the smallest incremental score of 1.
If the difference was equal or greater than 1, it would
have both statistical and clinical significance. We performed a post hoc sensitivity analysis. Using the current
cutoff score of 26 or greater as being a cognitive normal
result, 28 of the 103 subjects scored below 26 and would
be considered as failing this screening test. If we added
0.5 to all their scores, the same 28 of the 103 subjects
would have failed. If we added 1.0 to all of their scores,
then eight of the 28 subjects would then be considered
within normal range; thus, only 20 of the 103 subjects
would fail the HI-MoCA screening test.
Participants performed statistically significantly
better on the memory section of the HI-MoCA (score:
3.74, SD 5 1.145) compared to the original MoCA,
t(102)5 22.895, P 5 .005. (score: 4.09, SD 5 1.142).
However, for the language subtests, the HI-MoCA score
was statistically significantly lower, t(102) 5 3.57, P 5
.001 (score: 2.42, SD 5 0.69) than the original MoCA
(score: 2.08, SD 0.087). All the subtest scores and P values are listed in Table II.
A two-way repeated measures ANOVA was run to
determine whether the order of the testing—original
MoCA (T1), then HI-MoCA (T2 vs. HI-MoCA [T1)], and
then original MoCA [T2]—would impact the scoring for
group 1. Analysis of the studentized residuals showed
that there was normality, as assessed by the ShapiroWilk test of normality with no outliers, as assessed by
no studentized residuals greater than 6 3 SDs. There
was sphericity for the interaction term, as assessed by
Mauchly’s test of sphericity (P > .05). There was a statistically significant interaction between order and time on
MoCA outcome, F(1, 101) 5 6.08, P 5 .015, partial g2 5
.057. Therefore, simple main effects were run. MoCA
scores were statistically higher different if subjects
Laryngoscope 127: 2017
received the HI-MoCA first and the original MoCA second, F(1,42) 5 11.695, P 5 .001, partial g2 5 .002, compared to if they received the original MoCA first and the
HI-MoCA second, F (1,59) 5 .003, P - 0.96, partial g2 5
0. See Table II for means and SDs.
There were 49 subjects in group 2 who passed all
three screening tests and completed the HI-MoCA twice,
approximately 6 months apart.
In group 2, the average HI-MoCA score when tested
for the first time (T1) was 26.184 (SD 5 2.579). Approximately 6 months later when re-tested (T2), the average
HI-MoCA score was 26.493 (SD 5 2.639). There was no
statistically significant difference between the HI-MoCA
scores between the two time points, t(48) 5 2.995, P 5
0.324. The test–retest reliability coefficient was 0.66,
indicating excellent test–retest reliability (Table 3).
Results: Comparing HI-MoCA Scores in Normal
Hearing (Group 1) and Severely HearingImpaired (Group 2) Groups
We compared HI-MoCA scores in both group 1 and
group 2 subjects, which then were further analyzed to
determine whether severe hearing loss could impact HIMoCA scores. Group 1 subjects performed the HI-MoCA
either at the initial test period (n 5 60) or at the 6month period (n 5 43). Their average HI-MoCA score
was 26.66 (SD 5 2.642, standard error [SE] mean 5
0.378). Group 2 subjects were tested with the HI-MoCA
twice. Because the test–retest reliability coefficient was
very high (0.66) and a significant memory learning effect
was ruled out, for this comparison we only used the HIMoCA score on the first test timepoint (T1). The HIMoCA score for group 2 subjects was 26.184 (SD 5
2.579, SE mean 5 0.368). There was no statistically significant difference in HI-MoCA scores between group 1
or group 2 subjects (t(48) 5 1.138, P 5 0.261).
Further subtest analyses did not reveal any statistically significant differences in scores on the HI-MoCA
between group 1 and group 2 subjects (Table 4).
DISCUSSION
The goal of the present study was to develop a visual, computer-administered version of the MoCA for use
Lin et al.: Development of Hearing-Impaired MoCA
S9
in hearing-impaired individuals to directly address the
recognized need for tests adapted to individuals with
sensory impairments.22,23 We believe that, with our
results, our HI-MoCA is equivalent to the MoCA in our
tested population.
Study Design Metrics
We recognize that the conversion of auditory instructions to visual instructions would involve different neural
pathways. Although a detailed neuroanatomical discussion on the differences between auditory versus visual
processing pathways is beyond the scope of this article, we
strongly believed that our modifications were minimal but
necessary to achieve our study objective. Even among
experts within the field of neuroscience, there still is
debate on the precise neural pathways involved with
instruction-based task sets, irrespective of whether they
are auditory or visual in nature.27 We also believed that
although the neural entry point with auditory and visual
instructions are unique, the pathways involved with processing the instructions, storing information, and ultimately performing the requested tasks all merge into common
pathways. Furthermore, if changes were dramatic, differences in scoring would become apparent with the testing
of our two distinct groups using these two versions of the
MoCA. As demonstrated in our results, only small differences were noted, and none were clinically significant.
We also considered how best to eliminate patients
with any symptoms of cognitive decline for inclusion into
either group 1 or 2. In the original validation article,
Naasreddine et al.20 used a very comprehensive psychometric battery of tests for this determination. We initially
considered performing concurrent formal neuropsychological testing, but this was not possible given that there are
no versions that can be administered to the hearing
impaired. Based on prior screening efforts by our group, a
brief historical intake form and the three questionnaires
was determined to be adequate. The inclusion of an activities of daily living questionnaire by proxy, which was completed by the caregiver, further decreased the likelihood
that patients with significant cognitive impairment and
functional decline would be recruited. The global MoCA
and HI-MoCA results indicate that the majority of our
subjects in this study are cognitively intact.
statistically significant, it is not clinically relevant. Clinical relevance would be for statistically significant differences of 1 or greater. Our post hoc sensitivity analyses
did not confirm a clinically significant change in the
number of subjects who would fail the HI-MoCA screening test.
For screening tests to be widely utilized and
adopted, the test and its scoring system must be simple,
quick, and reliable. Although our HI-MoCA is arguably
more complex than the original MoCA, it remains an
easy-to-administer and reliable cognitive screening test.
As clinicians accustomed to working with the hearing
impaired well know, any form of interaction necessitates
an adapted means of communication, which is timeconsuming.
Consideration of Learning Effect and Test–
Retest Reliability
We also considered the possibility of whether subjects would improve their MoCA scores when being tested for the second time. We decided that a 6-month
window would be sufficiently long for subjects not to
demonstrate any significant memory effects. We also
were cognizant of the fact that a longer duration may
increase the window in which new health issues change
the subject’s cognitive state. Furthermore, a longer
follow-up period also could result in a higher dropout
rate among our subjects.
Fortunately, in our group 2 (severely hearing
impaired, normal cognition), the HI-MoCA scores at T1
and T2 were similar and not statistically significantly
different (26.66 vs. 26.18, P 5 0.261). The test–retest
reliability coefficient was 0.66.
Had we encountered a memory learning effect in
which the T2 scores were significantly improved, we
would have used alternative versions of the MoCA to
convert into several HI-MoCA versions as well. The original MoCA has a second and third version in which the
subtests are different. For example, rather than identify
the standard 3 animals, version 2 has a giraffe, bear,
and rhinoceros. The word list, number sequence, language, and abstraction examples also are different. For
simplicity, we elected to stay with one version.
Consideration of Test Order
Consideration of a Correction Factor
In our validation study, we found that in group 1
subjects (normal hearing, normal cognition), there was a
statistically significant difference in scores for the HIMoCA versus the MoCA (26.66 vs. 27.14). This difference
in score was 0.48. Further subtest analysis revealed that
two subtests of the seven were statistically significantly
different between the two versions of the MoCA (memory and language). However, the actual mean subtest difference ranged between 0.35 and 0.34. We considered a
correction factor to account for this difference. However,
the MoCA is scored with whole numbers with the smallest incremental score of 1. Although a 0.48 difference in
score between the HI-MoCA versus the MoCA might be
Laryngoscope 127: 2017
S10
We also asked whether the first test, regardless of
whether it was the HI-MoCA or MoCA, would factor in
the subjects’ scores. This was analyzed using the group
1 subjects only because they received both tests. A twoway repeated measures ANOVA analysis of both sequences (MoCA [T1] and then HI-MoCA [T2] vs. HI-MoCA
[T1] and then MoCA [T2]) did reveal that those in group
1 who were first evaluated with the HI-MoCA did worse
on the HI-MoCA than those who were first tested with
the MoCA.
This difference was a challenge to fully understand.
Those who were first tested with the HI-MoCA had an
average score of 26.12 and then scored 27.23 on the original MoCA, for a difference of greater than 1 point.
Lin et al.: Development of Hearing-Impaired MoCA
When the sequence of tests was reversed, the participants scored 27.07 and then scored 27.05 on the HIMoCA. We also note that the group 2 subjects who had
the HI-MoCA administered twice did not have any significant difference from T1 to T2 on their scores. Given
that this improvement in scoring was only seen in this
one instance, we decided it was not evidence of a learning effect. Again, this difference is not clinically significant because the clinical utility of this test is designed
for a single time point and not meant to be repeated.
Future Directions
We also have begun to test our HI-MoCA on a group
of subjects with clinically confirmed MCI. The purpose
of this study is to ensure that our HI-MoCA test results
also coincide with MoCA scores in patients who score
predominantly below the cutoff range of 26 and who
complete the necessary spectrum of patients of varying
cognitive functioning to fully validate the HI-MoCA.
With this set of data, we will be able to either confirm
26 or potentially a lower threshold in which patients
would be sent for further cognitive evaluations.
Our future direction also will include a smallerscale validation study using the original MoCA version 2
and 3, as well as a direct French HI-MoCA translation.
We also are in consideration of converting the timed
PowerPoint (Microsoft Corp., Redmond, WA) HI-MoCA
presentation into a multi OS mobile application (iOS
and Android). This mobile application, if run on a tablet,
could serve as both the testing platform and the recording and scoring platform, and it may allow worldwide
collaboration and the collection of data for future studies
and projects.
With the completion of this study, the HI-MoCA is
being utilized at our tertiary adult cochlear implant clinic to screen-selected patients over the age of 60 years
who may have signs of cognitive decline. We have also
distributed this test to various centers and colleagues in
North America, Europe, and Australia.
CONCLUSION
The HI-MoCA is an easy-to-administer and reliable
screening test for cognitive impairment in the severely
hearing impaired, with no conversion factor required in
our tested cohort of cognitively intact subjects.
Acknowledgment
We wish to thank Dr. Tony Esklander and Dr. Jane Opie
for assistance and advice on statistical analyses and article
preparation. We would also like to thank Dr. Ziad Nasreddine MD for his permission to modify the MoCA to the HIMoCA for this project.
Laryngoscope 127: 2017
BIBLIOGRAPHY
1. Uhlmann RF, Larson EB, Rees TS, Koepsell TD, Duckert LG. Relationship
of hearing impairment to dementia and cognitive dysfunction in older
adults. JAMA 1989;261:1916–1919.
2. Lin FR, Ferrucci L, Metter EJ, An Y, Zonderman AB, Resnick SM. Hearing loss and cognition in the Baltimore Longitudinal Study of Aging.
Neuropsychology 2011;25:763–770.
3. Lin FR, Metter EJ, O’Brien RJ, Resnick SM, Zonderman AB, Ferrucci L.
Hearing loss and incident dementia. Arch Neurol 2011;68:214–220.
4. Wong SL, Gilmour H, Ramage-Morin PL. Alzheimer’s disease and other
dementias in Canada. Health Rep 2016;27:11–16.
5. T Ostbye Crosse E. Net economic costs of dementia in Canada. CMAJ
1994;151:1457–1464.
6. Tam-Tham H, Nettel-Aguirre A, Silvius J, et al. Provision of dementiarelated services in Canada: a comparative study. BMC Health Serv Res
2016;16:184.
7. Gaugler JE, Westra BL, Kane RL. Professional discipline and support
recommendations for family caregivers of persons with dementia. Int
Psychogeriatr 2016;28:1029–1040.
8. Costa-Requena G, Espinosa Val MC, Cristofol R. Caregiver burden in
end-of-life care: advanced cancer and final stage of dementia. Palliat
Support Care 2015;13:583–589.
9. Richardson TJ, Lee SJ, Berg-Weger M, Grossberg GT. Caregiver health:
health of caregivers of Alzheimer’s and other dementia patients. Curr
Psychiatry Rep 2013;15:367.
10. Pattanayak RD, Jena R, Tripathi M, Khandelwal SK. Assessment of burden in caregivers of Alzheimer’s disease from India. Asian J Psychiatr
2010;3:112–116.
11. American Psychiatric Association. Diagnostic and Statistical Manual of
Mental Disorders, 5th ed. Arlington, VA: American Psychiatric Publishing; 2013.
12. McKhann GM, Knopman DS, Chertkow H, et al. The diagnosis of dementia due to Alzheimer’s disease: recommendations from the National
Institute on Aging-Alzheimer‘s Association workgroups on diagnostic
guidelines for Alzheimer’s disease. Alzheimers Dement 2011;7:263–269.
13. Ngo J, Holroyd-Leduc JM. Systematic review of recent dementia practice
guidelines. Age Ageing 2015;44:25–33.
14. Albert MS, DeKosky ST, Dickson D, et al. The diagnosis of mild cognitive
impairment due to Alzheimer’s disease: recommendations from the
National Institute on Aging-Alzheimer’s Association workgroups on
diagnostic guidelines for Alzheimer’s disease. Alzheimers Dement 2011;
7:270–279.
15. Langa KM, Levine DA. The diagnosis and management of mild cognitive
impairment: a clinical review. JAMA 2014;312:2551–2561.
16. Bondi MW, Smith GE. Mild cognitive impairment: a concept and diagnostic entity in need of input from neuropsychology. J Int Neuropsychol Soc
2014;20:129–134.
17. Brown DS, Bernstein IH, McClintock SM, et al. Use of the Montreal Cognitive Assessment and Alzheimer’s disease-8 as cognitive screening measures in Parkinson’s disease. Int J Geriatr Psychiatry 2016;31:264–272.
18. Saczynski JS, Inouye SK, Guess J, et al. The Montreal Cognitive Assessment: creating a crosswalk with the mini-mental state examination.
J Am Geriatr Soc. 2015;63:2370–2374.
19. Hoops S, Nazem S, Siderowf AD, et al. Validity of the MoCA and MMSE
in the detection of MCI and dementia in Parkinson disease. Neurology
2009;73:1738–1745.
20. Nasreddine ZS, Phillips NA, Bedirian V, et al. The Montreal Cognitive
Assessment, MoCA: a brief screening tool for mild cognitive impairment.
J Am Geriatr Soc 2005;53:695–699.
21. Zadikoff C, Fox SH, Tang-Wai DF, et al. A comparison of the mini mental
state exam to the Montreal cognitive assessment in identifying cognitive
deficits in Parkinson’s disease. Mov Disord 2008;23:297–299.
22. Dupuis K, Pichora-Fuller MK, Chasteen AL, Marchuk V, Singh G, Smith
SL. Effects of hearing and vision impairments on the Montreal Cognitive Assessment. Neuropsychol Dev Cogn B Aging Neuropsychol Cogn
2015;22:413–437.
23. Hill-Briggs F, Dial JG, Morere DA, Joyce A. Neuropsychological assessment of persons with physical disability, visual impairment or blindness,
and hearing impairment or deafness. Arch Clin Neuropsychol 2007;22:
389–404.
24. Hu C-J, Liao C-C, Chang C-C, Wu C-H, Chen T-L. Postoperative adverse
outcomes in surgical patients with dementia: a retrospective cohort
study. World J Surg 2012;36:2051–2058.
25. Gelina I, Gauthier L, McIntyre M. Development of a functional measure
for persons with Alzheimer’s disease: the Disability Assessment for
Dementia. Am J Occupational Therapy 1999;53:471–481.
26. Masellis M, Black SE. Assessing patients complaining of memory impairment. Geriatr Aging 2008;11:168–178.
27. Muhle-Karbe PS, Duncan J, De Baene W, Mitchell DJ, Brass M. Neural
coding for instruction-based task sets in human frontoparietal and visual cortex. Cereb Cortex 2016. doi:10.1093/cercor/bhw032.
Lin et al.: Development of Hearing-Impaired MoCA
S11