Behavioural Science Section

Gerontology 2009;55:570–581 Received: April 9, 2009

Accepted: April 9, 2009
DOI: 10.1159/000228918
Published online: July 15, 2009

Functional Abilities in Older Adults with

Mild Cognitive Impairment
Catherine L. Burton a Esther Strauss a David Bunce b Michael A. Hunter a David F. Hultsch a
a
Department of Psychology, University of Victoria, Victoria, B.C., Canada; b Centre for Cognition and Neuroimaging,
Brunel University, London, UK

Key Words older adults with no cognitive impairment on the EPT and
Alzheimer’s disease ⴢ Amnesia ⴢ Cognitive function ⴢ the SIB-R (both self- and informant-report versions). The
Everyday problem solving ⴢ Instrumental activities of daily multiple-domain MCI participants also demonstrated poor-
living ⴢ Mild cognitive impairment ⴢ Multiple-domain er IADLs than MCI participants with impairments in a single
impairment cognitive domain on the self-reported SIB-R and EPT. The
single-domain MCI groups demonstrated poorer IADLs than
older adults without cognitive impairment on the infor-
Abstract mant-reported SIB-R and EPT. No significant group differ-
Background: A classification scheme and general set of cri- ences were found on the Lawton and Brody IADL Scale. Us-
teria for diagnosing mild cognitive impairment (MCI) were ing the EPT and SIB-R as predictors in a multinomial regres-
recently proposed by a multidisciplinary group of experts sion analysis, MCI group status was reliably predicted, but
who met at an international symposium on MCI. One of the the classification rate was poor. Conclusion: Individuals
proposed criteria included preserved basic activities of daily with MCI demonstrated poorer IADL functioning compared
living and minimal impairment in complex instrumental to cognitively intact older adults. However, the changes in
activities of daily living (IADLs). Objective: To investigate IADL functioning observed in MCI may be too subtle to be
whether older adults with MCI classified according to the detected by certain measures, such as the Lawton and Brody
subtypes identified by the Working Group (i.e. amnestic, sin- IADL Scale. Copyright © 2009 S. Karger AG, Basel
gle non-memory domain, and multiple domain with or with-
out a memory component) differed from cognitively intact
older adults on a variety of measures indexing IADLs and to
examine how well measures of IADL predict concurrent MCI Instrumental Activities of Daily Living and Everyday
status. Methods: Two hundred and fifty community-dwell- Problem Solving in Older Adults with Mild Cognitive
ing older adults, ranging in age from 66 to 92, completed Impairment
self-report measures of IADLs (Lawton and Brody IADL Scale,
Scales of Independent Behaviour-Revised – SIB-R) and a Although interest in studying the intermediate state
measure of everyday problem solving indexing IADLs (Ev- between normal, healthy cognition and dementia has in-
eryday Problems Test – EPT). Ratings of participants’ IADL creased substantially over the past decade, much contro-
functioning were also obtained from informants (e.g. spouse,
adult child and friend). Results: Older adults with multiple-
domain MCI demonstrated poorer IADL functioning than Esther Strauss passed away on June 17, 2009.
198.143.54.65 - 1/27/2016 7:16:10 AM

© 2009 S. Karger AG, Basel David F. Hultsch

0304–324X/09/0555–0570$26.00/0 Department of Psychology, University of Victoria
Chalmers University of

Fax +41 61 306 12 34 PO Box 3050

E-Mail [email protected] Accessible online at: Victoria, B.C. V8W 3P5 (Canada)
Downloaded by:

www.karger.com www.karger.com/ger Tel. +1 250 721 6108, Fax +1 250 721 8929, E-Mail [email protected]
versy surrounds this area of research. Interest in this state, executive function and visuospatial skills (multiple-
often referred to as mild cognitive impairment (MCI), domain MCI), and (3) MCI with impairment in a sin-
originated from a desire to identify individuals at risk for gle non-memory domain (single non-memory-domain
developing Alzheimer’s disease (AD). Thus, conceptually, MCI). The multiple-domain MCI is further subdivided
the term MCI is often used to refer to a transitional zone into those with a memory impairment (multiple-domain
between normal cognitive function and early AD, or a MCI with amnesia) and those without (multiple-domain
preclinical state of AD [1, 2]. However, studies examining MCI without amnesia). The different subtypes likely re-
conversion rates from MCI to AD have shown that, al- flect different aetiologies, including degenerative, vascu-
though individuals with MCI have a greater risk of pro- lar, psychiatric or trauma-related causes [2, 8, 20]. For
gressing to AD than normal older adults, not all individ- example, the amnestic subtype of MCI is thought to pri-
uals with MCI go on to develop AD [3–5]. For example, marily reflect a prodromal phase of AD [21], whereas in-
in a population-based sample of older adults, Tuokko et dividuals with single non-memory domain MCI may
al. [6] reported that 32% of individuals initially diagnosed have a higher likelihood of progression to non-AD forms
with cognitive impairment but no dementia were diag- of dementia, such as dementia with Lewy bodies or fron-
nosed with AD 5 years later while 15% were diagnosed totemporal dementia.
with other forms of dementia (e.g. vascular dementia or In September 2003, a multidisciplinary, international
Parkinson’s disease). In fact, some individuals with MCI group of experts on MCI met to discuss the current status
appear to recover with time. Fisk and Rockwood [7] found and future directions of MCI research [1]. The Working
that 20–30% of individuals with MCI and objective mem- Group also endorsed a conceptualization of MCI that ac-
ory impairment showed no cognitive impairment at fol- knowledged the heterogeneity of the clinical presentation
low-up 5 years later. These results have led some research- and potential aetiologies associated with MCI. They pro-
ers to suggest that MCI represents a heterogeneous disor- posed a general set of criteria for diagnosing MCI, which
der with multiple possible outcomes [8–10]. include (1) absence of dementia (defined by DSM-IV or
Differences in the conceptualization of this transi- ICD 10); (2) cognitive decline [defined by (a) self and/or
tional state between normal aging and dementia are re- informant report in addition to evidence of objective cog-
flected in the varying terms and criteria used to opera- nitive impairment and/or (b) decline on objective mea-
tionalize it. Some of the most commonly used terms have sures of cognitive functioning], and (3) preserved basic
included age-associated cognitive decline [11], cognitive activities of daily living (ADLs) and minimal impairment
impairment, no dementia [12, 13], MCI [14] and mild in complex instrumental functions. Once a diagnosis of
neurocognitive disorder [15], all of which are associated MCI is achieved, the clinical presentation can then be
with varying inclusion and exclusion criteria. Although classified according to one of three subtypes: amnestic,
all of these concepts share a common focus (i.e. identify- multiple domain (with or without a memory component)
ing cognitive impairment in individuals who do not meet and single non-memory domain.
criteria for a diagnosis of dementia) [16], differences in As the criteria for MCI have evolved from the original
criteria are important because they impact estimates of criteria of Petersen et al. [14] to the current criteria estab-
prevalence, incidence and conversion rates [17–19]. lished by the Working Group [1], a subtle change in word-
One particularly well-known classification system ing is evident in the criterion pertaining to functional
was developed by Petersen et al. [14], referred to as MCI. abilities. Their original criteria for MCI specified ‘normal
It was originally proposed to identify individuals at risk activities of daily living’ [p. 304, 14], whereas the revised
for developing AD and the criteria included (1) the pres- version is somewhat less stringent and specifies ‘essen-
ence of a memory complaint, (2) normal activities of dai- tially’ or ‘largely’ intact functional activities, allowing for
ly living, (3) normal general cognitive functioning, (4) ‘minor inconveniences in daily function’ due to a cogni-
memory impairment relative to age peers and (5) the ab- tive impairment [p. 189, 20]. The Winblad et al. [1] criteria
sence of dementia. Later, in recognition of the heteroge- require preserved basic ADLs but allow for minimal im-
neous clinical presentation of MCI, the original criteria pairment in complex instrumental ADLs (IADLs). The
were revised to include other types of cognitive impair- report provided by the Working Group also suggests that
ment beyond memory [2]. The most recent classification future research is required to determine which specific
[8, 20] identifies three MCI subtypes: (1) MCI with a domains of IADLs might be impaired in MCI, thus ac-
memory impairment (MCI amnestic); MCI with impair- knowledging the possibility that individuals with MCI
ment in multiple cognitive domains, such as language, may show mild impairments in IADLs.
198.143.54.65 - 1/27/2016 7:16:10 AM

Functional Abilities and MCI Gerontology 2009;55:570–581 571

Chalmers University of
Downloaded by:
 The issue of whether functional abilities should be in- ficient sensitivity to detect subtle changes in functional
tact for a diagnosis of MCI has been an area of contention abilities. For instance, Artero et al. [27] argued that the
[9, 22]. Some researchers have challenged the criterion of measure of ADLs/IADLs used in their study contained
preserved everyday activities, citing recent evidence sug- finer gradations of change than conventional measures of
gesting that individuals with MCI may experience mild functional abilities, thus rendering it more sensitive to
functional impairments [19, 23]. For example, in a study minor changes in functional abilities. Others have ar-
comparing 107 individuals who met criteria for MCI gued, however, that individuals with MCI who have mild
according to the Clinical Dementia Rating Scale (i.e. levels of functional impairment may simply have more
CDR = 0 or 0.5) to older adults without cognitive impair- severe cognitive impairments and thus have mild demen-
ment, Tabert et al. [24] found that the MCI group self-re- tia [9, 10].
ported more difficulty with IADLs, as measured by the Given the recent development in the new MCI classi-
Pfeffer Functional Activities Questionnaire [25] and the fication scheme [1] and recent findings suggesting that
Lawton and Brody IADL Scale [26]. Artero et al. [27] ex- individuals with MCI experience subtle changes in their
amined functional abilities in a group of 308 older adults ability to perform everyday activities, we were interested
with MCI using a measure indexing both ADLs and in examining how these new MCI subtypes relate to func-
IADLs [Echelle Comportement et Adaptation Scale, 28]. tional abilities. Specifically, do older adults with MCI, as
Their criteria for MCI included proxy-reported changes defined by these new criteria, demonstrate greater im-
in cognitive performance over the past year (unrelated to pairment in functional abilities than older adults with no
physical illness), subjective cognitive complaints, absence cognitive impairment? In addition, do the MCI subtypes,
of dementia, and objective impairment in cognitive per- which differ in terms of the specific domains of cognitive
formance (i.e. ^1 SD) for age and education. They found functioning affected, also differ with respect to their per-
that 30.8% of their participants experienced difficulty formance on measures indexing IADLs? That is, are any
with at least one activity and that, over a 3-year period, of the MCI subtype patients particularly vulnerable to
declines in cognitive functioning were paralleled by loss- experiencing impairments in IADLs?
es in functional ability. Various methods have been developed for assessing
Similarly, Tuokko et al. [6] compared ADL and IADL functional abilities (e.g. self-report, informant-report, ob-
functioning, using the Older Americans Resources and servational, paper and pencil instruments), but currently
Services Scale [29], in a group of 712 older adults with no there is no gold standard. Research to date suggests that
cognitive impairment to a group of 225 individuals iden- different methods of assessing functional status provide
tified as having MCI without dementia (MCI/CIND) ac- different estimates of an individual’s ability to perform
cording to proxy report and clinical consensus. A larger IADLs [24, 30–32]. Consequently, it is possible that rela-
percentage of individuals in the MCI/CIND group dem- tionships between functional abilities and MCI status will
onstrated functional impairment (69.2%) than those differ depending on the type of IADL measures used. For
without cognitive impairment (44.9%), both with respect the present study, the relationship between MCI status and
to global functional ability and specific ADLs/IADLs (i.e. functional abilities was examined in a group of commu-
walking, showering and all of the IADL tasks), suggesting nity-dwelling older adults using self-report, informant-re-
that individuals with MCI/CIND may show significant port, and paper and pencil measures indexing IADLs.
functional disability. In an examination of the temporal A second purpose of the present study was to examine
relationships between functional and cognitive impair- whether the measures of IADL functioning could predict
ments across a 5-year interval, Tuokko et al. [6] also con- concurrent MCI status and if so, to evaluate how well the
cluded that deficits in the two areas tended to occur si- model classified participants.
multaneously.
Nygård [23] has argued that subtle changes, either ex-
perienced or observed, in complex everyday life activities Method
should be included in MCI criteria. She suggested that
explorative and qualitative studies of individuals with Participants
cognitive impairment or mild dementia indicate that The cross-sectional data for this study were obtained from the
3rd year of a longitudinal study (one of the IADL scales [26] was
these individuals experience subtle changes in IADLs in not added to the test battery until this year of the study). Of the
the early phases of cognitive deterioration. She also ar- original 304 community-dwelling older adults who completed
gued that traditional measures of IADLs might lack suf- the 1st year of the study, 250 participants (175 females and 75
198.143.54.65 - 1/27/2016 7:16:10 AM

572 Gerontology 2009;55:570–581 Burton /Strauss /Bunce /Hunter /Hultsch

Chalmers University of
Downloaded by:
males) ranging in age from 66 to 92 years (mean, M = 75.64, SD = ency and vocabulary) from an independent sample of 445 adults
5.69) completed the 3rd year of testing. Participants completed an aged 65–94 years recruited from the same population through the
average of 15.20 years of education (SD = 5.70, range 7–24) and all Victoria Longitudinal Study [40]1. Although published norms are
were Caucasian. available for most of these measures, they derive from a variety
of different samples with varying demographic characteristics.
Procedure Thus, the use of local norms derived for all tasks on the same
Participants were recruited for the study through local media population is preferred given the close demographic match to the
(newspaper and radio advertisements) requesting healthy com- current sample and the ability to make more accurate compari-
munity-dwelling volunteers who were concerned about their cog- sons across tasks.
nitive functioning. Potential participants were initially screened The normative sample was partitioned into four age by educa-
for exclusion criteria by a telephone interview. Exclusionary cri- tion groups (age = 65–74 years and 75+ years; education = 0–12
teria included a diagnosis of dementia by their physician or a Mini years and 13+ years). Means and standard deviations were com-
Mental State Examination (MMSE) [33] ! 24, a history of signifi- puted for these groups for the five measures and these normative
cant head injury (defined as loss of consciousness for more than values were used to classify participants from the present sample
5 min), other neurological or major medical illnesses (e.g. heart on the basis of cognitive status. Given that specific cut-off scores
disease, cancer and Parkinson’s disease), severe sensory impair- have not been specified by Petersen [20] or Winblad et al. [1] for
ment, extensive drug or alcohol abuse, current psychiatric diag- identifying MCI status and following the 1 SD criterion used in
noses or psychotropic drug usage. studies conducted by Artero et al. [27] and Ritchie et al. [19], par-
Individuals identified as potential participants attended a ticipants were classified as MCI if they obtained scores 11.0 SD
group testing session to assess cognitive functioning in five do- below the mean of their age- and education-matched peers from
mains: perceptual speed [WAIS-R Digit Symbol Substitution, 34], the normative sample on the cognitive reference tasks. The MCI
inductive reasoning [Letter Series, 35], episodic memory [Imme- participants were then further subcategorized into the MCI sub-
diate Free Word Recall, 36], verbal fluency [Controlled Associa- types based on the particular areas of cognitive impairment dem-
tions, 37] and vocabulary [Extended Range Vocabulary, 37]. Par- onstrated: amnestic MCI (n = 6), multiple-domain MCI (with am-
ticipants also provided information regarding demographic char- nesia: n = 19; without amnesia: n = 28) and single non-memory
acteristics and self-ratings of health during the group testing domain MCI (n = 39). The remaining participants were classified
session. as not cognitively impaired (NCI; n = 158).
Following the group testing session, an individual intake in- Table 1 shows the age, education and benchmark cognitive
terview was conducted with each participant, at which time they status of the participants at year 3 of the study. One-way ANOVAs
completed a series of benchmark cognitive measures such as the conducted on these variables revealed significant group dif-
Wechsler Adult Intelligence Scale-III (WAIS-III) Block Design ferences for age [F (4, 245) = 21.67, p ! 0.001], education [F (4,
and Vocabulary subtests [38] and the North American Adult 245) = 20.04, p ! 0.001], MMSE [F (4, 245) = 5.61, p ! 0.001], Block
Reading Test [NAART, 39]. Estimates of current general intellec- Design [F (4, 245) = 7.81, p ! 0.001], Vocabulary [F (4, 245) = 9.78,
tual status (full-scale IQ or FSIQ) were computed based on the p ! 0.001], WAIS-III Estimated FSIQ [F (4, 245) = 13.57, p !
age-adjusted Block Design and Vocabulary subtest scaled scores 0.001] and NAART Estimated IQ [F (4, 245) = 13.93, p ! 0.001].
[40], and estimates of premorbid intellectual status were based on Post hoc contrasts using Tukey’s HSD for unequal sample sizes
the NAART [39]. (p ! 0.05) indicated that the NCI group was younger than all of the
Following the intake interview, participants were assessed on MCI groups while the multiple-domain MCI with amnesia group
five separate sessions, approximately biweekly. During each ses- was older than the single non-memory domain MCI group. The
sion, participants completed a battery of cognitive tasks (e.g. read- NCI group was also more highly educated than the single non-
ing test and episodic memory) and state-like indicators of physi- memory domain MCI, multiple-domain MCI with amnesia and
cal (e.g. gait, blood pressure and respiratory functioning) and multiple-domain MCI without amnesia groups, with the single
emotional functioning (e.g. beliefs and affect). Participants and non-memory domain MCI group also being more highly educated
designated informants also completed measures of functional than the multiple-domain MCI with amnesia group. With respect
abilities once per year. For each of the following waves of testing to the MMSE, the amnestic MCI group obtained lower scores than
(i.e. years 2–4), the structure of the sessions remained the same. the NCI, single non-memory domain MCI and multiple-domain
Participants were categorized into groups on the basis of cog- MCI without amnesia groups. On Block Design, the NCI group
nitive status using the consensus classification scheme by Win- performed better than the single non-memory domain MCI, mul-
blad et al. [1]. Classifying participants on the basis of cognitive tiple-domain MCI with amnesia and multiple-domain MCI with-
status is ideally done using norms from a set of reference cognitive out amnesia groups. The NCI group also performed better on Vo-
tasks obtained from a separate sample. Normative data were cabulary than the multiple-domain MCI with and without amnesia
available for the five indicators of cognitive functioning (i.e. per- groups. In addition, the single non-memory domain MCI outper-
ceptual speed, inductive reasoning, episodic memory, verbal flu- formed the multiple-domain MCI with amnesia group on Vocabu-
lary. The NCI group demonstrated higher estimated current and
premorbid intellectual abilities (i.e. WAIS-III Estimated FSIQ and
1 NAART Estimated IQ) than the single non-memory domain MCI,
Data on all 445 participants were available for the measures of percep-
tual speed, reasoning, verbal fluency, and vocabulary, but, due to imple- multiple-domain MCI with amnesia and multiple-domain MCI
mentation of a counterbalancing procedure, information on the episodic without amnesia groups. In addition, the multiple-domain MCI
memory task was only available for 194 of the 445 participants of the nor- with amnesia group had a lower NAART Estimated IQ than the
mative sample. amnestic MCI and single non-memory domain MCI groups.
198.143.54.65 - 1/27/2016 7:16:10 AM

Functional Abilities and MCI Gerontology 2009;55:570–581 573

Chalmers University of
Downloaded by:
Table 1. Demographic characteristics as a function of cognitive status

NCI A-MCI Non-A MCI MD MCI + A MD MCI – A

Number 158 6 39 19 28
Age, years 73.5784.72 79.5085.65 77.5485.61 82.0085.04 79.5784.86
Years of education 16.2982.70 13.6783.93 14.2382.67 12.0082.71 12.7982.71
MMSE 28.9281.17 26.8382.48 28.6781.26 28.1681.26 28.6881.09
WAIS-III Block Designa 13.5982.85 13.8381.60 11.8782.78 11.7481.82 11.1482.66
WAIS-III Estimated FSIQ 125.10811.22 125.0086.75 117.51811.42 111.7989.01 113.3689.51
Estimated NAART IQ 118.5685.00 118.0583.48 115.2687.11 110.1588.81 112.3487.31

NCI = Not cognitively impaired; A-MCI = amnestic MCI; non-A MCI = single non-memory domain MCI; MD MCI + A = mul-
tiple-domain MCI with amnesia; MD MCI – A = multiple-domain MCI without amnesia.
a WAIS-III Age-Scaled Scores.

Measures likely that the omission of these earlier items had little impact on
Self-Reported IADLs overall estimates of functional independence. The internal consis-
Participants completed two self-report questionnaires assess- tency (Cronbach’s ␣) for the modified SIB-R for year 1 of the cur-
ing their capacity to perform a variety of IADLs. The Lawton and rent study was 0.92. One-year test-retest data were available for 210
Brody IADL scale [26] is a self-report measure indexing eight participants; the test-retest reliability was r = 0.80.
IADL domains: telephone use, food preparation, housekeeping,
laundry, transportation, shopping, financial management and Informant-Reported IADLs
medication use. Participants rated their level of independence in The Lawton and Brody IADL scale and the SIB-R were also
each of the eight IADL domains according to Likert scales rang- administered to individuals identified by each participant as a
ing from 0 to 2–4. The items were summed to provide a total score significant other. Informants willing to complete the measures
(total possible score = 31), with lower scores indicating higher lev- were available for 230 of the participants and included spouses
els of functional independence. (36.8%), friends (19.4%), adult children (15.8%), siblings (0.7%)
The Scales for Independent Behavior-Revised (SIB-R) [42] is a and other relationships (3.0%). The informants ranged in age
measure of functional independence, and adaptive and problem from 29 to 90 (M = 66.71, SD = 13.11) and had an average of 15.34
behaviours. The functional independence and adaptive behav- years of education (SD = 3.13, range 7–27). Most of the informants
iour component is comprised of 13 subscales organized into four had known their participant for 120 years (75.7%), while 8.7%
adaptive behaviour clusters: Motor (gross- and fine-motor skills), had known them for 10–20 years, 8.7% for 5–10 years, 5.2% for
Social Interaction and Communication (social interaction, lan- 3–5 years and 1.7% for ^2 years. The frequency of contact varied:
guage comprehension and language expression), Personal Living 38.2% daily and living in same household, 12.5% daily or almost
(eating and meal preparation, toileting, dressing, personal self- daily but living in separate households, 17.8% one to two times a
care and domestic skills) and Community Living (time and punc- week and 5.6% one to two times a month (1.6% unknown fre-
tuality, money and value, work skills and home/community ori- quency of contact).
entation). Each subscale contains items that span the entire devel- The internal consistency (Cronbach’s ␣) and test-retest reli-
opmental range, from early infancy to 80 years and older. Items ability coefficient for the informant-report SIB-R were 0.95 and
within each subscale are arranged in approximate developmental r = 0.84, respectively.
order, from the easiest to the hardest. For each item, individuals
are asked how well they are able to complete the task (or would be Paper and Pencil Measure
able to complete the task if they do not typically perform it), with- The Everyday Problems Test (EPT) [43] is a paper and pencil
out help or supervision, using a 4-point rating scale, ranging from measure of everyday cognitive competence that requires partici-
0 (never or rarely performs the task, even if asked) to 3 (almost pants to solve problems associated with daily living. Participants
always does the task very well without being asked). Raw scores are provided with 21 printed stimulus materials (e.g. medication
were calculated for each adaptive behaviour subscale by summing label and pay phone information) and asked to solve two problems
the ratings of all items, with higher scores indicating higher levels pertaining to each stimulus. All stimuli are reproductions of real-
of functional independence. life materials with four to six items representing one of the follow-
The SIB-R was originally designed as an informant-reported ing IADL domains: medication use, meal preparation, telephone
measure. However, for the purposes of the present study, it was ad- use, shopping, financial management, household management and
ministered directly to participants to provide a self-reported mea- transportation. For example, with respect to meal preparation, par-
sure of functional abilities. In addition, items pertaining to early ticipants are provided with a nutritional information chart for cere-
infancy and early childhood were omitted in order to shorten the als and are asked to determine how many calories are in a serving
length of time required for administration. As all of the partici- of cereal if whole milk is used instead of skim milk. For a question
pants were still relatively highly functioning (i.e. no dementia), it is in the transportation domain, participants are provided with a
198.143.54.65 - 1/27/2016 7:16:10 AM

574 Gerontology 2009;55:570–581 Burton /Strauss /Bunce /Hunter /Hultsch

Chalmers University of
Downloaded by:
chart of taxi rates and are asked to determine how much they would Table 2. Correlations between the measures of functional abilities
have to pay if they traveled 1 mile in a suburban area. The original and age and education
EPT consists of 42 items; however, for the purposes of the present
study, items 10 (Medicare Benefits Payment Schedule) and 12 (Stain Variable Age Education
Removal Directions) were omitted because a significant minority
of participants disputed interpretation of elements of the problems. 1 Self-reported L & B IADL 0.24*** –0.10
Consequently, possible scores ranged from 0 to 40, with higher 2 Informant-reported L & B IADL 0.22** –0.10
scores indicating better performance. 3 Self-reported SIB-R –0.31*** 0.15*
Willis and Marsiske [43] reported a 1-year test-retest reliabil- 4 Informant-reported SIB-R –0.32*** 0.17*
ity of 0.93. The internal consistency (Cronbach’s ␣) for the EPT 5 EPT –0.41*** 0.32***
for year 1 of the current study was 0.88. With respect to validity,
the EPT has been found to correlate significantly (r = 0.67) with L & B IADL = Lawton and Brody IADL Scale.
direct observation of older adults’ performance of everyday tasks * p < 0.05; ** p < 0.01; *** p < 0.001.
involving medications, meal preparation and phone usage [43].
Correlations with older adults’ self-reports (r = 0.23) and demen-
tia patients’ self-reports (r = 0.36) of IADL functioning were also
significant, but lower [43, 44].
[Levene’s F = 6.44, p = 0.012; t (39.42) = –2.50, p = 0.017]
than those who had missing data for one or more mea-
Results sures. They did not differ with respect to age or education
or on the year-1 scores for the EPT, WAIS-III Vocabulary,
Sample Attrition WAIS-III Estimated IQ or NAART Estimated IQ.
In the 1st year of the study, 304 older adults participat-
ed in the study (208 females and 96 males). Of the partici- Group Differences in IADL Measures
pants who did not return for the 3rd year of the study, 10% In the first set of analyses, group differences in the
were deceased; 22% withdrew because of health problems; IADL measures were examined using one-way ANOVAs,
6% withdrew because of memory problems; 6% withdrew covarying age and education. Not only were there signifi-
for family reasons (e.g. decline in spouse’s health); 8% had cant group differences for age and education, but, as
moved, and 48% were too busy or not interested in con- shown in table 2, most of the IADL measures were also
tinuing. Participants who returned at year 3 did not differ significantly correlated with age and education. In addi-
significantly from those who did not with respect to edu- tion, given the large number of potential post hoc com-
cational level, WAIS-III Vocabulary or NAART Estimated parisons, we elected to examine group differences using
IQ. However, participants who returned for year 3 of the planned contrasts (p ! 0.05) comparing: (1) the NCI group
study were younger [dropouts: M = 75.76, SD = 6.79; re- to the single-domain MCI groups (i.e. amnestic MCI and
turnees: M = 73.64; SD = 5.69; t (302) = 2.39, p = 0.018] and single non-memory domain MCI); (2) the NCI group to
they obtained higher scores at year 1 on the EPT [drop- the multiple-domain MCI groups (i.e. multiple-domain
outs: M = 26.79, SD = 8.17; returnees: M = 31.54; SD = 5.61; MCI with and without amnesia), and (3) the single-do-
t (302) = –5.17, p ! 0.001]; MMSE [dropouts: M = 28.32, main MCI groups to the multiple-domain MCI groups.
SD = 1.51; returnees: M = 28.88, SD = 1.10; t (289) = –2.85, Group differences in performance on the five IADL
p = 0.005]; WAIS-III Block Design [dropouts: M = 10.76, measures are shown in figures 1–5. Significant group
SD = 2.64; returnees: M = 12.51; SD = 2.78; t (289) = –3.78, differences emerged for the self-reported SIB-R 2 [F (4,
p ! 0.001], and WAIS-III Estimated FSIQ [dropouts: M = 242) = 4.78, p ! 0.001], the informant-reported SIB-R
113.93, SD = 12.30; returnees: M = 121.17; SD = 12.39; [F (4, 219) = 3.26, p = 0.013] and the EPT [F (4, 241) = 14.5,
t (289) = –3.47, p = 0.001]. p ! 0.001]. For the self-reported SIB-R, the NCI group
and the single-domain MCI groups reported better IADL
Missing Data functioning than the multiple-domain MCI groups [NCI:
In contrast to the informant-reported measures, the F (1, 242) = 17.83, p ! 0.001; single-domain MCI groups:
percentage of missing data was small for the EPT (0.8%), [F (1, 242) = 4.46, p = 0.036]. For the informant-reported
self-reported IADLs (0.4%) and self-reported SIB-R (0.4%).
Using independent-sample t tests and Levene’s test for
Equality of Variances, participants with data for all five 2
One-way ANOVAs were also run for each of the SIB-R subscales sepa-
measures obtained higher scores at year 1 on the MMSE rately and a similar pattern of results was obtained as when using the total
[t (248) = –2.04, p = 0.042] and WAIS-III Block Design scores for the self-reported and informant-reported SIB-R scales.
198.143.54.65 - 1/27/2016 7:16:10 AM

Functional Abilities and MCI Gerontology 2009;55:570–581 575

Chalmers University of
Downloaded by:
 Color version available online
2.5

2.0

Total score
1.5

1.0

0.5

0
NCI Amnestic MCI Multiple- Multiple- Single
domain MCI domain MCI non-memory
with amnesia without domain MCI
amnesia

Color version available online

2.5

2.0
Total score

1.5

1.0

0.5

0
NCI Amnestic MCI Multiple- Multiple- Single
domain MCI domain MCI non-memory
with amnesia without domain MCI
Fig. 1, 2. Self-reported (1) and informant- amnesia
reported Lawton and Brody IADLs (2) as a
function of cognitive status. 2

SIB-R, the NCI group was rated by their significant oth- higher intercorrelations with one another than with the
ers as having significantly better functional abilities than EPT.
the single-domain MCI groups [F (1, 219) = 7.40, p =
0.007] and multiple-domain MCI groups [F (1, 219) = Prediction of Group Membership
9.93, p = 0.002]. On the EPT, the NCI group obtained sig- A multinomial logistic regression analysis was per-
nificantly better scores than the single-domain MCI formed in order to examine whether the IADL variables
groups [F (1, 241) = 8.58, p = 0.004], who in turn obtained could predict concurrent MCI group status (i.e. MCI
significantly better scores than the multiple-domain MCI group status at year 3). Only those measures that showed
groups [F (1, 241) = 53.4, p ! 0.001]. significant group differences in the previous analyses
Given that the MCI groups also differed with respect (i.e. self-reported SIB-R, informant-reported SIB-R and
to NAART Estimated IQ, the analyses were also conduct- EPT) were used as predictor variables. A total of 224 par-
ed using NAART Estimated IQ as a covariate, along with ticipants had data for all three IADL measures. The sin-
age and education, and the same pattern of results was gle-domain MCI groups were collapsed into one group
obtained. (n = 43) and the multiple-domain impairment groups
Table 3 shows the correlations between the various were collapsed into another (n = 39).
IADL measures, using Pearson correlation coefficients. A test of the full model (i.e. with all 3 predictors) showed
The Lawton and Brody IADL and SIB-R measures show that the predictors, as a group, reliably predicted group
198.143.54.65 - 1/27/2016 7:16:10 AM

576 Gerontology 2009;55:570–581 Burton /Strauss /Bunce /Hunter /Hultsch

Chalmers University of
Downloaded by:
 Color version available online
220
200
180
160
140

Total score
120
100
80
60
40
20
0
NCI Amnestic MCI Multiple- Multiple- Single
domain MCI domain MCI non-memory
with amnesia without domain MCI
amnesia
3

Color version available online

220
200
180
160
140
Total score

120
100
80
60
40
20
0
NCI Amnestic MCI Multiple- Multiple- Single
domain MCI domain MCI non-memory
Fig. 3, 4. Self-reported (3) and informant- with amnesia without domain MCI
reported SIB-R (4) as a function of cogni- amnesia
tive status. 4

Color version available online

40
35
30
Total score

25
20
15
10
5
0
NCI Amnestic MCI Multiple- Multiple- Single
domain MCI domain MCI non-memory
with amnesia without domain MCI
amnesia
Fig. 5. Performance on the EPT as a func-
tion of cognitive status.
198.143.54.65 - 1/27/2016 7:16:10 AM

Functional Abilities and MCI Gerontology 2009;55:570–581 577

Chalmers University of
Downloaded by:
Table 3. Intercorrelations between the
measures of functional abilities Variable 1 2 3 4 5

1 Self-reported L & B IADL –

2 Informant-reported L & B IADL 0.69** –
3 Self-reported SIB-R –0.58** –0.49** –
4 Informant-reported SIB-R –0.57** –0.69** 0.62** –
5 EPT –0.23** –0.19** 0.33** 0.36** –

L & B IADL = Lawton and Brody IADL Scale. ** p < 0.01.

membership [␹2 (6) = 96.40, p ! 0.001]. In addition, using Table 4. Adequacy of classification based on the logistic regres-
a deviance criterion, the model was not reliably different sion analysis model
from a perfect model, indicating that the predictors pro-
Observed Predicted
vided a good fit [␹2 (78) = 41.15, p = 1.00]. According to
the Nagelkerke R 2, the set of predictors accounted for NCI single- multiple-
domain MCI domain MCI
41.7% of the variance in group membership. The model
was significantly degraded by the removal of the self-re- NCI 137 0 5
ported SIB-R [␹2 (2) = 6.23, p = 0.044] and EPT [␹2 (2) = Single-domain MCI 30 5 8
56.91, p ! 0.001], suggesting that only these two pre- Multiple-domain MCI 12 4 23
dictors reliably distinguished between the MCI status
groups. The Wald statistic indicated that both the self-re-
ported SIB-R [Wald (1) = 37.74, p ! 0.001] and the EPT
[Wald (1) = 6.07, p = 0.014] reliably separated the multiple- types. Thus, this study represents the first study to exam-
domain MCI group from the NCI group, but that only the ine functional abilities in the different MCI subtypes.
EPT [Wald (1) = 18.81, p ! 0.001] reliably separated the In the present community-dwelling sample, approxi-
single-domain MCI group from the NCI group. mately one third of the participants (36.8%) were classi-
As shown in table 4, classification was excellent for the fied into one of the MCI subtypes. Amongst the MCI par-
NCI group (96.5% correct) but much poorer for the mul- ticipants, the most common classification was the single
tiple-domain MCI (59.0% correct) and the single-domain non-memory domain MCI subtype, with 42.4% of the
MCI groups (11.6%), for an overall correct classification MCI participants falling into this category. Approximate-
rate of 73.7%. Clearly, participants were overclassified in ly one half of the MCI participants were classified as one
the largest group: the NCI group. of the multiple-domain MCI subtypes (20.7% with mem-
ory impairment and 30.4% without memory impairment).
Only 6.5% of the total MCI sample was classified as the
Discussion amnestic subtype. Although the overall prevalence rate of
MCI in the present study is substantially higher than
The primary purpose of the present study was to ex- prevalence rates reported by previous studies [17, 19, 45],
tend previous research examining the relationship be- it is important to bear in mind that most studies have ex-
tween MCI and functional abilities by examining the re- amined the amnestic type of MCI. Thus, the current prev-
lationship between the Winblad et al. [1] MCI classifica- alence rate of 2.4% obtained for the amnestic MCI subtype
tion scheme and ability to perform IADLs in a group of is similar to estimated prevalence rates of 3.2 and 3.7%
community-dwelling older adults. The original MCI cri- reported by Ritchie et al. [19] and Kumar et al. [17], respec-
teria of Petersen et al. [14] focused on memory impair- tively, for amnestic MCI. Fisk et al. [45] found that the
ment with relative preservation of other cognitive do- prevalence rate for the amnestic subtype varied depend-
mains, and, thus, most research to date is based on this ing on the restrictiveness of the criteria used; they found
more narrow conceptualization of MCI. A broader con- that the prevalence rate increased from 1.03 to 3.02%
ceptualization of MCI has been adopted in recent years, when the requirements for subjective memory complaints
recognizing other areas of cognitive impairment [1, 20], and intact IADLs were eliminated. In contrast to the pres-
but, as of yet, little is known about these other MCI sub- ent study, which used psychometric cut-off points and a
198.143.54.65 - 1/27/2016 7:16:10 AM

578 Gerontology 2009;55:570–581 Burton /Strauss /Bunce /Hunter /Hultsch

Chalmers University of
Downloaded by:
relatively liberal criterion (i.e. 1 SD), case assignment in with functional abilities, suggesting that these individu-
the study by Fisk et al. [45] was based on clinical judge- als are especially likely to require additional assistance
ment. Of interest is the similarity in prevalence rates ob- and support in performing IADLs.
tained in these various studies, despite the use of differing The present results raise the question of why individu-
criteria used to classify MCI participants. als with multiple-domain MCI demonstrate poorer per-
To date, the only other study that has examined the formance on measures of IADL functioning than indi-
prevalence rates of the different MCI subtypes was con- viduals with single-domain impairments, as found on the
ducted by Rasquin et al. [46]. They too found that the EPT and the self-reported SIB-R. It is possible, as has
amnestic subtype was the least common subtype in a been suggested by Strauss et al. [47] and Davis and Rock-
memory clinic sample (14.4%) and a cohort of first-time wood [10], that greater impairment in functional abilities
stroke patients (0%). However, the prevalence of the oth- simply reflects more severe cognitive impairment, and
er subtypes differed depending on the sample. For the thus more advanced disease progression. Rasquin et al.
memory clinic sample, Rasquin et al. [46] reported that [46] also proposed that individuals with multiple-do-
the multiple-domain, amnesia subtype was the most main MCI are in the predementia stage and have a high
common (44.1%), followed by the single non-memory do- risk of developing dementia. However, examining par-
main (22%) and multiple-domain impairment without ticipants’ performance on the MMSE, Block Design, Vo-
amnesia (19.5%) subtypes. For the stroke cohort, on the cabulary and WAIS-III Estimated FSIQ does not shed
other hand, the multiple-domain impairment without much light on this issue. Participants with multiple-do-
amnesia group was the most prevalent (57.5%), followed main MCI with amnesia performed more poorly relative
by the single non-memory domain (26.2%) and multiple- to the single-domain MCI groups on some of the cogni-
domain impairment with amnesia (16.3%) groups. Thus, tive tasks (i.e. WAIS-III Vocabulary and NAART Esti-
clearly, estimated prevalence rates for the MCI subtypes mated IQ) but not all (i.e. MMSE, WAIS-III Estimated
differ depending on the specific population from which FSIQ and Block Design), and no significant differences
samples are drawn. However, based on the preliminary were obtained between the multiple-domain MCI with-
results provided by their study and the current study, it out amnesia and the single-domain MCI groups. Thus,
appears as though the amnestic MCI subtype may be rel- the pattern of results is inconsistent, perhaps reflecting a
atively uncommon regardless of sampling. lack of power due to widely discrepant samples sizes. An
One of the aims of the current investigation was to ex- alternative explanation may be that some specific types
amine whether individuals with MCI show differential of aetiologies associated with multiple-domain MCI are
performance on measures indexing IADLs. The results more likely to affect functional abilities (e.g. frontotem-
of the present study indicated that individuals with im- poral lobe dementia) compared to the types of aetiologies
pairments in multiple cognitive domains (with or with- typically associated with single-domain MCI.
out amnesia) demonstrated poorer IADL functioning Three different measures of IADL functioning were
than participants without cognitive impairment (i.e. used in the present study, two of which had both self-re-
NCI) on three of the five IADL measures, namely the self- ported and informant-reported versions. In line with past
reported SIB-R, the informant-reported SIB-R and the research demonstrating that information derived from
EPT. The multiple-domain MCI participants also dem- measures of functional status varies depending on the
onstrated poorer IADL functioning than individuals method of assessment [24, 30–32], the current study found
with impairments in a single cognitive domain (i.e. am- that the pattern of results differed depending on the par-
nestic MCI and single non-memory domain MCI) on the ticular measure of IADL examined. Although group dif-
self-reported SIB-R and the EPT. Finally, significant dif- ferences were observed for the SIB-R and the EPT, no sig-
ferences between the NCI group and the single-domain nificant group differences were found on one of the most
groups were obtained for the informant-reported SIB-R commonly used measures of IADLs (i.e. Lawton & Brody
and the EPT, with the latter demonstrating poorer IADL IADL Scale) [26], suggesting that it may lack sufficient sen-
functioning than the former. Thus, the results of the pres- sitivity to detect the subtle changes in functional abilities
ent study suggest that older adults with MCI do show de- observed in MCI. The SIB-R and EPT contain finer grada-
clines in IADL performance compared to cognitively in- tions of change and may be more sensitive to minor chang-
tact older adults. In addition, it appears as though indi- es in functional abilities. However, despite tapping into
viduals with impairments in multiple cognitive domains IADLs, the EPT is not actually a direct measure of IADLs
are particularly vulnerable to experiencing difficulties but, rather, was designed as a measure of everyday cogni-
198.143.54.65 - 1/27/2016 7:16:10 AM

Functional Abilities and MCI Gerontology 2009;55:570–581 579

Chalmers University of
Downloaded by:
tive competence or everyday problem solving [43]. There- likely reflect true differences in the prevalence rates for
fore, the EPT may be reflecting changes in cognitive func- each of the subtypes, future research would benefit from
tioning more so than changes in functional abilities, as larger sample sizes, allowing for a more accurate repre-
demonstrated by the small- to moderate-sized correlations sentation of the more uncommon MCI subtypes, espe-
with the Lawton and Brody and SIB-R measures. However, cially the amnestic MCI group.
if cognitive impairments drive the simultaneous decline in In summary, the present study demonstrated that,
functional abilities observed in individuals with MCI, as when using a fairly liberal psychometric definition of
suggested by Tuokko et al. [6], it would be expected that cognitive impairment (i.e. !1 SD), MCI is quite common
decrements in performance on the EPT would also be par- in a community-dwelling sample of older adults, with al-
alleled by decrements on cognitively complex IADLs. most half of the sample meeting criteria for one of the
An additional aim of the present study was to examine MCI subtypes. Although the debate as to whether to in-
whether MCI group status could be predicted by concur- clude or exclude functional impairment for identifying
rent performance on the IADL variables. Only the EPT individuals with MCI continues, the present study dem-
and the SIB-R, both self- and informant-reported ver- onstrates that individuals with MCI do demonstrate im-
sions, demonstrated significant group differences and paired IADL functioning compared to cognitively intact
were therefore used as predictors of MCI group status. A older adults and that those with cognitive impairment in
multinomial logistic regression analysis revealed that, al- multiple areas may be particularly vulnerable to demon-
though the predictors as a group reliably predicted group strating impairments in functional abilities. In addition,
status, the classification rate for the single-domain MCI certain measures of IADL functioning appear to be more
group was very poor. These results suggest that measures sensitive to detecting subtle differences in IADL func-
of functional abilities contribute important information tioning (e.g. SIB-R and EPT) than others (e.g. Lawton and
in predicting MCI group status, but alone, they are insuf- Brody IADL).
ficient in terms of distinguishing between the different
subtypes.
There are several potential limitations associated with Acknowledgements
this study that need to be mentioned. Participants were
not randomly selected from the community, resulting in This research was supported by grants from the Medical Re-
a sample of well-educated, Caucasian, older adults; thus, search Council of Canada, the Canadian Institutes of Health Re-
the generalizability of the present results is limited. How- search and the Alzheimer Society of Canada to David F. Hultsch,
Esther Strauss and Michael A. Hunter, and a doctoral award from
ever, given that participants lost to attrition demonstrat- the Alzheimer Society of Canada Institute of Aging/Canadian In-
ed poorer initial levels of cognitive functioning and that stitutes of Health Research and a Senior Graduate Studentship
significant group differences in functional abilities were trainee award from the Michael Smith Foundation for Health Re-
found in the remaining, high-functioning sample at year search to Catherine L. Burton. Preparation of this article was also
3, it is possible that even stronger group differences may supported by a Research Unit Infrastructure Grant from the Mi-
chael Smith Foundation for Health Research to the University of
have been found with a more varied, representative sam- Victoria Centre on Aging. David Bunce’s collaboration in this
ple. Secondly, the classification scheme was applied to the project was supported by a grant from the Association of Com-
sample following data collection using psychometric cut- monwealth Universities/British Academy.
off scores. Thus, classification was not made on an indi- We thank Dr. Roger Dixon for the use of data from the Victo-
vidual basis, using clinical judgement as specified by Pe- ria Longitudinal Study to establish norms for the classification of
cognitive status of the present sample.
tersen [20] and Winblad et al. [1], and therefore, the pres-
ent determination of MCI status does not represent a
clinical diagnosis. In addition, it cannot be ascertained
whether areas of cognitive impairment reflect an actual References 1 Winblad B, Palmer K, Kivipelto M, Jelic V,
Fratiglioni L, Wahlund LO, Nordberg A,
decline in cognitive performance. Similarly, it is not pos- Bäckman L, Albert M, Almkvist O, Arai H,
sible to determine if the difficulties with functional abil- Basun H, Blennow K, De Leon M, Decarli C,
ities manifested by participants are due to cognitive im- Erkinjuntti T, Giacobini E, Graff C, Hardy
JCJ, Jorm AF, Ritchie K, Van Duijn C, Visser
pairment as opposed to medical comorbidities or physi- PJ, Petersen RC: Mild cognitive impairment
cal limitations. Finally, the number of participants in – beyond controversies, towards a consen-
sus: report of the International Working
each MCI group was widely discrepant, complicating Group on Mild Cognitive Impairment. J In-
statistical analyses. Although differences in sample size tern Med 2004;256:240–246.
198.143.54.65 - 1/27/2016 7:16:10 AM

580 Gerontology 2009;55:570–581 Burton /Strauss /Bunce /Hunter /Hultsch

Chalmers University of
Downloaded by:
 2 Petersen RC, Doody R, Kurz A, Mohs RC, 17 Kumar R, Dear KBG, Christensen H, llsch- 32 Myers AM, Holliday PJ, Harvey KA,
Morris JC, Rabins PV, Ritchie K, Rossor M, ner S, Jorm AF, Meslin C, Rosenman SJ, Hutchinson KS: Functional performance
Thal L, Winblad B: Current concepts in mild Sachdev PS: Prevalence of mild cognitive measures: are they superior to self-assess-
cognitive impairment. Acta Neurol Scand impairment in 60- to 64-year-old communi- ments? J Gerontol 1993; 48:M196–M206.
2001;58:1985–1992. ty-dwelling individuals: the Personality and 33 Folstein M, Folstein S, McHugh PR: Mini-
3 Bruscoli M, Lovestone S: Is MCI really just Total Health through Life 60+ Study. Dement Mental State: a practical method for grading
early dementia? A systematic review of con- Geriatr Cogn Disord 2005;19:67–74. the cognitive state of patients for the clini-
version studies. Int Psychogeriatr 2004; 16: 18 Bischkopf J, Busse A, Angermeyer MC: Mild cian. J Psychiatr Res 1975;12:189–198.
129–140. cognitive impairment: a review of preva- 34 Wechsler D: Manual for the Wechsler Adult
4 Morris JC, Storandt M, Miller JP, McKeel lence, incidence and outcome according to Intelligence Scale-Revised. New York, Psy-
DW, Price JL, Rubin EH, Berg L: Mild cogni- current approaches. Acta Psychiatr Scand chological Corporation, 1981.
tive impairment represents early-stage Alz- 2002;106:403–414. 35 Thurstone RG: Primary Mental Abilities:
heimer disease. Arch Neurol 2001; 58: 397– 19 Ritchie K, Artero S, Touchon J: Classification Grades 9–12, revised. Chicago, Science Re-
405. criteria for mild cognitive impairment: a search Associates, 1962.
5 Comijs HC, Dik MG, Deeg DJH, Jonker C: population-based validation study. Neurol- 36 Hultsch DF, Hertzog C, Dixon RA: Ability
The course of cognitive decline in older per- ogy 2001;56:37–42. correlates of memory performance in adult-
sons: results from the Longitudinal Aging 20 Petersen RC: Mild cognitive impairment as a hood and aging. Psychol Aging 1990; 5: 356–
Study Amsterdam. Dement Geriatr Cogn diagnostic entity. J Intern Med 2004; 256: 368.
Disord 2004;17:136–142. 183–194. 37 Ekstrom RB, French JW, Harman HH, Der-
6 Tuokko H, Morris C, Ebert P: Mild cognitive 21 Morris JC: Mild cognitive impairment is ear- men D: Manual for Kit of Factor Referenced
impairment and everyday functioning in ly-stage Alzheimer disease. Arch Neurol Cognitive Tests. Princeton, Educational
older adults. Neurocase 2005; 11:40–47. 2006;63:15–16. Testing Service, 1976.
7 Fisk JD, Rockwood K: Outcomes of incident 22 Rivas-Vazquez RA, Mendez C, Rey GJ, Car- 38 Psychological Corporation: WAIS-III Tech-
mild cognitive impairment in relation to razana EJ: Mild cognitive impairment: new nical Manual. San Antonio, Psychological
case definition. J Neurol Neurosurg Psychia- neuropsychological and pharmacological Corporation, 1997.
try 2005;76:1175–1177. target. Arch Clin Neuropsychol 2004;19:11– 39 Blair JR, Spreen O: Predicting premorbid IQ:
8 Petersen RC: Conceptual overview; in Pe- 27. a revision of the National Adult Reading
tersen RC (ed): Mild Cognitive Impairment: 23 Nygård L: Instrumental activities of daily Test. Clin Neuropsychol 1989;3:129–136.
Aging to Alzheimer’s Disease. Oxford, Ox- living: a stepping-stone towards Alzheimer’s 40 Dixon RA, de Frias C: The Victoria Longitu-
ford University Press, 2003, pp 1–14. disease diagnosis in subjects with mild cog- dinal Study: from characterizing cognitive
9 Collie A, Maruff P: An analysis of systems of nitive impairment? Acta Neurol Scand 2003; aging to illustrating changes in memory
classifying mild cognitive impairment in 107:42–46. compensation. Aging Neuropsychol Cogn
older people. Aust NZ J Psychiatry 2002; 36: 24 Tabert MH, Albert SM, Borukhova-Milov L, 2004;11:346–376.
133–140. Camacho Y, Pelton G, Liu X, Stern Y, Deva- 41 Sattler JM, Ryan JJ: Assessment of Children:
10 Davis HS, Rockwood K: Conceptualization nand DP: Functional deficits in patients with Revised and Updated Third Edition WAIS-
of mild cognitive impairment: a review. Int J mild cognitive impairment: prediction of III Supplement. La Mesa, Sattler, 1999.
Geriatr Psychiatry 2004;19:313–319. AD. Neurology 2002;58:758–764. 42 Bruininks RH, Woodcock RW, Weatherman
11 Levy R: Aging-associated cognitive decline. 25 Pfeffer RI: Measurement of functional activ- RF, Bradley BK: Scales of Independent Be-
Int Psychogeriatr 1994;6:63–68. ities in older adults in the community. J havior-Revised: Comprehensive Manual.
12 Ebly EM, Hogan DB, Parhad IM: Cognitive Gerontol 1982;37:323–329. Itasca, Riverside Publishing, 1996.
impairment in the non-demented elderly: re- 26 Lawton MP, Brody E: Assessment of older 43 Willis SL, Marsiske M: Manual for the
sults from the Canadian Study on Health people: self-maintaining and instrumental Everyday Problems Test. University Park,
and Aging. Arch Neurol 1995;52:612–619. activities of daily living. Gerontologist 1969; Pennsylvania State University, 1993.
13 Tuokko H, Frerichs RJ: Cognitive impair- 9:179–186. 44 Willis SL, Allen-Burge R, Dolan MM, Ber-
ment with no dementia (cind): longitudinal 27 Artero S, Touchon J, Ritchie K: Disability trand RM, Yesavage J, Taylor JL: Everyday
studies, the findings, and the issues. Clin and mild cognitive impairment: a longitudi- problem solving among individuals with
Neuropsychol 2000; 14:504–525. nal population-based study. Int J Geriatr Alzheimer’s disease. Gerontologist 1998; 38:
14 Petersen RC, Smith GE, Waring SC, Ivnik RJ, Psychiatry 2001;16:1092–1097. 569–577.
Tangalos EG, Kokmen E: Mild cognitive im- 28 Ritchie K, Ledésert B: The measurement of 45 Fisk JD, Merry HR, Rockwood K: Variations
pairment: clinical characterization and out- incapacity in the severely demented elderly: in case definition affect prevalence but not
come. Arch Neurol 1999;56:303–308. the validation of a behavioural assessment outcomes of mild cognitive impairment.
15 American Psychiatric Association: Diagnos- scale. Int J Geriatr Psychiatry 1991; 6: 217– Neurology 2003; 61:1179–1184.
tic and Statistical Manual of Mental Disor- 226. 46 Rasquin SMC, Lodder J, Visser PJ, Lousberg
ders, ed 4. Washington, American Psychiat- 29 Fillenbaum GG: Multidimensional Func- R, Verhey FRJ: Predictive accuracy of MCI
ric Association, 2000. tional Assessment of Older Adults: The Duke subtypes for Alzheimer’s disease and vascu-
16 Palmer K, Fratiglioni L, Winblad B: What is Older Americans Resources and Services lar dementia in subjects with mild cognitive
mild cognitive impairment? Variations in Procedures. Hillsdale, Erlbaum, 1988. impairment: a 2-year follow-up study. De-
definitions and evolution of nondemented 30 Østbye R, Tyas S, McDowell K, Koval J: Re- ment Geriatr Cogn Disord 2005;19:113–119.
persons with cognitive impairment. Acta ported activities of daily living: agreement 47 Strauss E, Slick DJ, Levy-Bencheton J, Hunt-
Neurol Scand 2003;107(suppl 179):14–20. between elderly subjects with and without er M, MacDonald SWS, Hultsch DF: Intrain-
dementia and their caregivers. Age Ageing dividual variability as an indicator of malin-
1997;26:99–106. gering in head injury. Arch Clin Neuropsy-
31 Loewenstein DA, Argueelles S, Bravo M, chol 2002;17:423–444.
Freeman RQ, Argueelles T, Acevedo A, Eis-
dorfer C: Caregivers’ judgments of the func-
tional abilities of the Alzheimer’s disease
patient: a comparison of proxy reports and
objective measures. J Gerontol B Psychol Sci
Soc Sci 2001;56:P78–P84.
198.143.54.65 - 1/27/2016 7:16:10 AM

Functional Abilities and MCI Gerontology 2009;55:570–581 581

Chalmers University of
Downloaded by: