ADCS-ADL Scale, Scoring and Manual - Dementia Outcomes ...

@]
Alzheimer's Disease
Cooperative Study
~@mJl]~~@ (g@~@ ~@~@[J11 [P@[JmJl]
ADCS - Activities of Daily Living Inventory Page 1 of 8
Informationobtained through: 0 Informant visit
0 Telephone call
1. Regardingeating:
Which best describes {S} usual performanceduring the past 4 weeks?
3 0 ate without physical help, and used a knife
20 used a fork or spoon, but not a knife, to eat
10
00
used fingers to eat
{S} usually or always was fed by someone else
2. Regardingwalking (or getting around in a wheelchair), in the past 4 weeks,
which best describes his/her optimal performance:
3 0 mobile outside of home without physical help
2 0 mobile across a room without physical help
1 0 transferred from bed to chair without help
0 0 required physical help to walk or transfer
3. Regarding bowel and bladder function at the toilet, which best describes
his/her usual performance in the past 4 weeks:
3 0 did everything necessary without supervision or help
2 0 needed supervision, but no physical help
1 0 needed physical help, and was usually continent
0 0 needed physical help, and was usually incontinent
Galasko, D., Bennett, D., Sano, M., Ernesto, E., Thomas, R., Grundman, M., and Ferris, S.
Alzheimer Disease and Associated Disorders 1997; 11:S33-S39.
Used with permission from the NIA Alzheimer's Disease Cooperative Study
ADL 1 .0 2/99

@] .@CID [j[Jj) [j2) ~ @ ~ CID
Alzheimer's Disease
Cooperative Study
Yes No
Don't
Know 6.
l
D
0
D
0
~
@ !JB @ [j2) @ [f It [F @ [f [j[Jj)
ADCS - Activities of Daily Living Inventory Page 2 of 8
4. Regardingbathing, in the past 4 weeks, which best describes his/her usual
performance:
3 D bathed without reminding or physical help
2 D no physical help, but needed supervision/reminders to bathe completely
1 D needed minor physical help (e.g., with washing hair) to bathe completely
0 D needed to be bathed completely
5. Regardinggrooming£ in the past 4 weeks, which best describes his/her optimal
performance:
3
2
1
0
D cleaned and cut fingernails without physical help
D brushed or combed hair without physical help
D kept face and hands clean without physical help
D needed help for grooming of hair, face, hands, and fingernails
Regarding dressing, in the past 4 weeks:
A) Did {S} select his/her first set of clothes for the day?
If yes, which best describes his/her usual performance:
~ 3 D without supervision or help
2 D with supervision
1 D with physical help
B) Regarding physically getting dressed, which best describes his/her usual
performance in the past 4 weeks:
4 D dressed completely without supervision or physical help
3 D dressed completely with supervision, but without help
2 D needed physical help only for buttons, clasps, or shoelaces
1 D dressed without help if clothes needed no fastening or buttoning
0 D always needed help, regardless of the type of clothing
2
ADL 1.0 2/99

@]
Alzheimer's Disease
Cooperative Study
Yes
L
Yes
L
No
Don't
Know
0
0
0
0
Don't
Yes No Know
1
0 0
0 0
01 01 0
0
0
0
01 0 0
No
Don't
Know
0
0
0
0
~@)ITlfl)~~@(G@)~@ [pJ@~@[f11[P@[fITlfl)
ADCS - Activities of Daily Living Inventory
7. In the past 4 weeks, did {8} use a telephone?
8.
Don't
Yes No Know 9.
0 0
0 0
If yes, which best describes his/her highest level of performance:
Page 3 of 8
;:. 5 D made calls after looking up numbers in white or yellow pages, or by
4
3
2
1
dialing directory assistance
D made calls only to well-known numbers, without referring to a directory
or list
D made calls only to well-known numbers, by using a directory or list
D answered the phone; did not make calls
D did not answer the phone, but spoke when put on the line
In the past 4 weeks, did {8} watch television?
If yes, ask all questions:
Did (8):
a) usually select or ask for different programs or his/her favorite show?
b) usually talk about the content of a program while watching it?
c) talk about the content of a program within a day (24 hours) after watching it?
In the past 4 weeks, did {8} ever appear to pay attention to conversation or small
talk for at least 5 minutes?
Note: {5} did not need to initiate the conversation.
If yes, which best describes his/her usual degree of participation:
3 D usually said things that were related to the topic
2 D usually said things that were not related to the topic
1 D rarely or never spoke
10. Did {8} clear the dishes from the table after a meal or snack?
If yes, which best describes how he/she usually performed:
;:. 3 D without supervision or help
2 D with supervision
1 D with physical help
3
ADL 1.0 2/99

~
Alzheimer's Disease
Cooperative Study
Yes
Yes
Yes
Don't
~@mru[9)~@ (g@~@ [g1@[9)@[{lt ~@[{mru
ADCS - Activities of Daily Living Inventory Page 4 of 8
No
D
0
Know
D
0
11. In the past 4 weeks, did {S} usually manage to find his/her personal belongings at
home?
If yes, which best describes how he/she usually performed:
Don't
No Know
D D 12.
0
~ 0
Yes
Don't
No Know 13.
D D
Don't
No Know 14
D D .
0
0
0
0
3 D without supervision or help
2 D with supervision
1 D with physical help
In the past 4 weeks, did {S} obtain a hot or cold beverage for him/herself?
(A cold beverage includes a glass of water.)
If yes, which describes his/her highest level of performance
3 D made a hot beverage, usually without physical help
2 D made a hot beverage, usually if someone else heated the water
1 D obtained a cold beverage, usually without physical help
In the past 4 weeks, did {S} make him/herself a meal or snack at home?
If yes, which best describes his/her highest level of food preparation:
~ 4 D cooked or microwavedfood, with little or no help
3 D cooked or microwaved food, with extensive help
2 D mixed or combined food items for a meal or snack, without cooking or
microwaving (e.g., made a sandwich)
1 D obtained food on his/her own, without mixing or cooking it
In the past 4 weeks, did {S} dispose of garbage or litter in an appropriate place
or container at home?
If yes, which best describes how he/she usually
3 D without supervision or help
2 D with supervision
1 D with physical help
4
performed:
ADL 1.0 2/99

@]
Alzheimer's Disease
Cooperative Study
Yes
L
Yes
D 1
Yes
Don't
No Know 15.
D D
0 0
Don't
No Know 16.
D D
0 0
D 0
D 0
Don't
No Know 17.
D D
0 0
~@lmru~~@(g@l~@ [Ri@~@[fli ~@[fmru
ADCS - Activities of Daily Living Inventory Page 5 of 8
In the past 4 weeks, did {S} get around (or travel) outside of his/her home?
If yes, which best describes his/her optimal performance:
;.. 4 D alone, went at least 1 mile away from home
3 D alone, but remained within 1 mile of home
2 D only when accompanied and supervised, regardless of the trip
1 D only with physical help, regardlessof the trip
In the past 4 weeks, did {S} ever go shopping?
If yes, ask A and B:
A) Which one best describes how {S} usually selects items:
3 D without supervision or physical help?
2 D with some supervision or physical help?
1 D not at all, or selected mainly random or inappropriate items?
B) Did {S} usually ~ for items without supervision or physical help?
In the past 4 weeks, did {S} keep appointments or meetings with other people,
such as relatives, a doctor, the hairdresser, etc.?
If yes, which best describes his/her awareness of the meeting ahead of time:
3 D usually remembered, may have needed written reminders
e.g. notes, a diary, or calendar
2 D only remembered the appointment after verbal reminders on the day
1 D usually did not remember, in spite of verbal reminders on the day
5
ADL 1.0 2/99

@]
Alzheimer's Disease
Cooperative Study
Yes No
Don't
Know 18.
0 0
? D D
D D D
1 0 0
D D D
1 0 0
D D D
1 0 0
D 1 D 0 D 0
D 1 D 0 D 0
Don't
Yes No Know 19.
D D
0 0
D D D
1 0 0
D D D
1 0 0
D D D
1 0 0
Don't
Yes No Know
D D 20.
0 0
~@QVi)~~@ (Q;@~@ ~@~@[f11 ~@[fQVi)
ADCS - Activities of Daily Living Inventory Page 6 of 8
In the past 4 weeks, was {S} ever left on his/her own?
If yes, ask all questions:
Was {S} left:
a) away from home, for 15 minutes or longer, during the day?
b) at home, for an hour or longer, during the day?
c) at home, for less than 1 hour, during the day?
In the past 4 weeks, did {S} talk about current events? (This means events or
incidents that occurred during the past month.)
If yes, ask all questions:
Did {S} talk about events that...:
a) he/she heard or read about or saw on TV but did not take part in?
b) he/she took part in outside home involving family, friends, or neighbors?
c) events that occurred at home that he/she took part in or watched?
In the past 4 weeks, did {S} read a magazine. newspaper or book for more than 5
minutes at a time?
If yes, ask all questions:
Did {S} usually:
a) talk about details of what he/she read while or shortly
« than 1 hour) after reading?
b) talk about what he/she read 1 hour or longer after reading?
6
ADL 1.0 2/99

~ ~@J[ffi1)~~@(g@J~@ [Ri@~@[flt[P@[f[ffi1)
Alzheimer's Disease
Cooperative Study
Yes
Yes
No
Don't
Know
D
0
D
0
Don't
No Know 22.
D D
0 0
ADCS - Activities of Daily Living Inventory Page 7 of 8
21. In the past 4 weeks, did {S} ever write things down?
Note: If {S} wrote things only after encouragement or with help, the response
should still be 'yes.'
If yes, which best describes the most complicated things that he/she wrote:
3 D letters or long notes that other people understood
2 D short notes or messages that other people understood
1 D his/her signature or name
In the past 4 weeks, did {S} perform a pastime. hobby or game?
If yes, which pastimes did he/she perform:
Ask about all of the following, check all that apply:
D card or board games (including bridge, chess, checkers)
D bingo D crosswords Dart
D musical instrument D knitting D sewing
Dreading D gardening D golf
D tennis
D other
D workshop D fishing
Note: Walking does NOT count as a hobby/pastime for this scale.
D If {S} performs hobbies/pastimes only at day care, check here.
If yes, how did {S} usually perform his/her most common pastimes:
3 D without supervision or help
2 D with supervision
1 D with help
7
ADL 1 .0 2/99

@I ~@)M~~@ ~@)~@ [Ri@~@[f1t [F@[fM
Alzheimer's Disease
Cooperative Study
Yes No
Don't
Know
D
0
D
0
L
ADCS - Activities of Daily Living Inventory Page 8 of 8
23. In the past 4 weeks, did {S} use a household appliance to do chores?
~ Ask about all of the following, and check those that were used:
D washer D dryer D vacuum
D dishwasher D toaster D toaster oven
D range D microwave D food processor
D other
If yes, for the most commonly used appliances, which best describes how {S}
usually used them:
4 D without help, operating more than on-off controls if needed
3 D without help, but operated only on/off controls
2 D with supervision, but no physical help
1 D with physical help
LU Total Score (0-78)
LU Number of "Don't Know" Responses
8
ADL1.0 2/99

1.1-'
ADCS - ACTIVITIES OF DAILY LIVING (ADL) INVENTORY
NOTES: (1) {P} refers to the participant and should be replaced by the participant's
name or relationship to the study partner each time an ADL question
is asked of the study partner.
(2) This ADL inventory must be given in the format of an interview of the
study partner, either directly or by telephone. The form should NOT be
given to a study partner to complete on his/her own.
READ THE FOllOWING INSTRUCTIONS TO THE STUDY PARTNER:
I am goingto ask you about a numberof daily activitiesthat {P} may have performed
duringthe past4 weeks. Pleasetell me about{P}'s actual performance,not aboutwhat
he/shecould havedone if an opportunityhad arisen. For each activitythat {P}
attempted,I'm going to ask you to chooseone out of a numberof descriptionsthat best
fits his/her most usual performance.
For some activities, I'll ask about whether {P} performed independently, or with
supervision or help. Let me explain how we are defining these words:
Independently means that {P} completed the activity without being helped. We still
consider it independent if {P} was reminded or prompted to get started, or
received a little prompting while performing the activity.
With supervision means that {P} required verbal reminders and instructions while
doing the activity.
With help means that {P} was given some degree of ~ assistance by another
person to perform the activity.
INSTRUCTIONS FOR THE RATER:
If the study partner states that {P} had no opportunity to perform the task during the past
four weeks (e.g., {P} did not have access to a telephone, therefore could not possibly
make phone calls), the response should be recorded as 'no.'
If either the study partner's answer or the questionnaire are unclear, please make notes
on the case report form detailing the problem.
For questions regarding specific ADL items, please refer to the ADL Response Card.
I-I

ADCS - Activitiesof Daily Living
(ADL) Inventory
A. General Remarks
There are widely varying ways to carry
out ADL, especially Instrumental ADL.
This leads to difficulty when trying to
obtain ADL ratings from an informant in a
standardized way for a clinical trial. The
ADCS - ADL Inventory approaches the
problem by offering detailed descriptions
of each activity, and by asking the
informant to describe observed actions or
behaviors. The informant is asked to
focus on the past 4 weeks. The informant
must not estimate what the patient might
be able to do had an opportunity arisen,
but on what the patient actually did. The
informant should not try to interpret the
patient's thought processes or intentions.
To help the informant to remain focused
on observed actions and behavior during
the past 4 weeks, it may be useful to ask
him/her for examples of what the patient
did regarding the ADL in question.
B. Administering the ADCS - ADL
Inventory
The ADCS - ADL Inventory was developed
and tested as an interview administered
by a rater in person or by telephone.
It should not be filled out by the
informant.
C. Format of Questions
. For each basicADL (questions 1-5,
6A), there is a forced choice of best
response.
. All other ADL consist of a main
.
question followed by subquestions
(descriptors) .
Subquestions are arranged in
hierarchical fashion, starting with the
highest (most independent) level of
.
.
.
ADL performance and ending with
the lowest.
For each ADL, the initial response to
the main questions is "yes", "no" or
"don't know". If an informant gives 4
or more "don't know" responses, it is
worth trying to identify an alternative
informant.
After a "no" or "don't know" response,
the subquestions are
disregarded unless specific instructions
indicate otherwise. "No" or
"don't know" act as fast forward cues
to proceed to the next ADL.
After a "yes" response, there are 2
possible paths:
1. From several descriptors, the
informant chooses the one that best
matches the patient's performance
(e.g., question 6A). The informant
should be offered as many descriptors
as necessary to describe the
patient's ADL ability, starting from the
highest level and proceeding downwards
or
2. The informant makes a "yes" or
"no" choice for each subquestion.
There is a reminder in these cases to
ask every question after an initial
"yes" (e.g., question 8).
D. Standardizing "independently",
"with supervision",
and "with help"
For many ADL, the hierarchy depends on
how much intervention is needed by the
informant or others to enable the patient
to perform the ADL.
Independently = the patient completed
the ADL without physical help, and at
most with reminders to do the task, or a
brief prompt during the ADL.

ADL
With supervision = the patient required
verbal reminders and instructions while
doing the ADL; this occupied the
caregiver's (or informant's) time.
With help = the patient was given some
degree of physical assistance by another
person to perform the ADL.
E. "Usual" performance of ADL
Patients may vary from day to day in their
ADL performance. If an informant
responds that a patient sometimes
performs ADL at one level and sometimes
at another, the interviewer should
ask him/her to choose the most com-
monly applicable level/descriptor.
When In doubt about a higher or lower
level of ability, rate the higher one if the
patient does manage to perform at that
level fairly consistently.
F. Comments on specific questions
in the Inventory
1. Eating: self-explanatory
2. Walking: we are interested in mobility,
not specifically walking. If the patient
used a walker or wheelchair, but was
mobile outside of home without help, he/
she is independent.
3. Toileting: self-explanatory
4. Bathing: minor physical help includes
actions such as washing hair, help with
drying, running the water or adjusting its
temperature. More extensive help should
be scored as needing to be bathed.
5. Grooming: toenail cutting is not rated
since physical difficulty may impede this
aspect of grooming, even in cognitively
normal elderly.
6A. "Selecting" clothes impliesactive
participationbythe patient.This may
involve physically selecting clothes, or
providing input to the caregiver about
wishes or preferences. It is rated separately
from physically getting dressed.
68. Dressing: the 3rd and 4th levels are
similar. The intention is that "help only for
buttons, clasps or shoelaces" implies
only a minor contribution from the
caregiver, i.e., the patient performs some
degree of fastening, zipping or tying.
Using "clothes needing no fastening or
buttoning" implies that these have been
completely abandoned. Velcro counts as
a fastener.
7. Telephone: reminders or supervision
are allowable but not help, e.g., a patient
can be prompted to dial directory assistance,
or told the number and still score
at the highest level. If the patient makes
calls only if the numbers are dictated by
the caregiver, or if the telephone is set up
to automatically dial one of a preselected
group of numbers on pressing a single
button, the patient scores 2 points.
8. Television: if the patient sits in front
of a television screen without demon-
strating awareness or recollection of
something he/she sees, then all
subquestions will be answered as "no".
"Talk about the content of a program"
should be interpreted fairly broadly; the
patient does not need to initiate the
conversation, but should require more
than a "yes" or "no" answer to a question
like "Did vou enjoy the program?"
9. Conversation: "paid attention"
implies more than being present and
seemingly alert during a conversation.
The patient must participate verbally.
Since it is difficult to judge whether
nonverbal participation in a conversation
is meaningful, it is not an option among
the descriptors.
10. Clear dishes from a table: the
patient does not have to clear the table
entirely, but should participate enouQ,hto

ADL
make a useful contribution. Clearing
items other than dishes e.g., glasses or
utensils are acceptable as alternatives.
11. Finds personal belongings: it may
be helpfulto give examplessuch as
clothing,glasses,wallet, keys,etc.
12. Beverage: gettingor preparinga
beverageafter instructionsor a remindercounts
as no physical help.
13. Mealor snack: if the patient
required supervision to cook or microwave,
but functioned without physical
help, score as the highest level. Mixing
or combining items without cooking can
also include supervision. An example of
the lowest level of a "yes" response is
the patient who finds and eats food preprepared
by someone else.
14. Dispose of garbage or litter: does
not only refer to major household
garbage produced in a kitchen. Disposing
of any trash in an appropriate
container qualifies for a "yes" response.
15. Travel: intended to cover the
patient's ability to remain oriented, not
get lost and be able to venture beyond
home. It does not matter whether the
patient walked, drove, took public
transport or was a passenger in a car.
The distance of 1 mile is arbitrary and
implies travel beyond sight of home.
16. Shopping: shopping is a complex
activity. We have focused on the two
most essential aspects (after getting to
the shop, covered by question 15),
choosing items and paying. If the patient
goes shopping with a prewritten list and
a sum of money provided by someone
else, that should be scored as without
supervision, provided the patient
selected the items on the list. Paying for
items could involve cash, check or
credit; the issue is whether supervision
or help is needed.
17. Keep appointments or meetings:
mainly aimed at monitoring memory. An
"appointment/meeting" may be liberally
interpreted to include almost any kind of
preplanned meeting, outing or excursion.
It does not matter who made the appointment
or whether the patient traveled
alone or with somebody else. The highest
level implies the patient remembered the
appointment on his/her own, memory
aids are acceptable.
18. Left alone: if the patient was left
alone for an hour or longer at home, then
they automatically will score a "yes" for
descriptor c, "less than 1 hour at home."
19. Talk about current events: the
patient does not need to initiate conversation
about current events, but must
demonstrate awareness or recollection
by providing details of the event(s).
Merely agreeing or disagreeing with other
people by saying "yes or "no" is not
sufficient to display knowledge/recall of
current events. Patients who are severely
aphasic will score "no" for talking about
current events.
20. Reading: looking at a book, magazine
or newspaper and turning the pages
for more than 5 minutes on end may
equal reading. Unless the patient communicates
the content of what he/she
reads to someone else, it is not possible
to judge whether he/she actually reads or
not, in a way that shows elements of
comprehension and retention of information.
The informant should be encouraged
to interact with the patient to be
able to make an accurate judgment.
21. Writing: short notes or messages
can be either spontaneous or written to
dictation. A shopping list, "to do" list, or
taking an intelligible telephone message
would qualify.
22. Pastime, hobbl.2!.Same: the

ADL
menu of hobbies or games is to help the
informant and may provide us with
secondary information if complex hobbies
are lost and replaced by simpler ones. A
hobby should involve element(s) of
concentration, knowledge and memory,
and manual skills. If hobbies other than
those in the menu are offered, try to
describe what the patient does in some
detail so that we can monitor this.
23. Household appliance: an appliance
is defined as a device with one or
more switches or controls, usually
(though not necessarily) with an electrical
or other power source, used to do
chores. For men, tools with controls or
switches would qualify, but not a screwdriver,
hammer or saw. Again, if the
patient needs only minimal verbal
prompting to operate the appliance,
press a switch, or operate a control, the
highest level is scored.

Alzheimer Disease and Associated Disorders
Vol. 11, Suppl. 2, S33-S39
@ 1997 Lippincott-Raven Publishers, Philadelphia
An Inventory to Assess Activities of Daily Living for Clinical
Trials in Alzheimer's Disease
)
Douglas Galasko, *David Bennett, tMary Sano, Chris Emesto, Ronald Thomas,
Michael Grundman, :j:Steven Ferris, and the Alzheimer's Disease Cooperative Study
Department of Neurosciences (Neurology), University of California, San Diego, California; *Rush Alzheimer's Disease Center
and Department of Neurological Sciences, Rush-Presbyterian-St. Lukes Medical Center, Chicago, Illinois; tDepartment of
Neurology, Columbia University and the Sergievsky Center, New York, New York; and :j:Departmentof Psychiatry, New York
University Medical Center, New York, New York, u.S.A.
Summary: We developed a set of infonnant-based items describing perfonnance of
activities of daily living (ADL) by patients with Alzheimer's disease (AD) to identify
which ADL are useful for assessment of patients in clinical trials. Evaluation of ADL
is an important outcome measure in AD clinical trials. For clinical trial measurement,
ADL should have broad applicability, good test-retest reliability, scaling to cover a
range of perfonnance, and sensitivy to detect change in disease progression. A total of
45 ADL items developed from literature review and clinical experience were administered
to informants of 242 AD patients and 64 elderly controls as part of the multicenter
Alzheimer's Disease Cooperative Study Instrument protocol. Half of the subjects
were re-evaluated at 1 and 2 months and all at 6 and 12 months. Controls
performed virtually all ADL items optimally at baseline and at 12 months. Among
subjects with AD, 27 of the 45 ADL were widely applicable, i.e., performed at baseline
or premorbidly by >90% of subjects; showed good test-retest reliability between
baseline and 1 and 2 months; correlated with MMSE scores of AD patients crosssectionally;
and showed a decline in performance from baseline to 12months in at least
20% of AD patients. ADL could be identified that capture change in functional ability
in patients across the entire range of the MMSE. The remaining 18 ADL included
several that may be useful for trials that target specific populations, e.g., women with
AD. Because change on specific items depends on baseline MMSE, ADL evaluation
should include items relevant to the severity of dementia of patients enrolled in a
clinical trial. Key Words: Alzheimer's disease-Activities of daily living-Clinical
trials-Patient assessment.
Functional assessment of patients with Alzheimer's
disease (AD), in terms of performance of activities of
daily living (ADL), is a critical element in patient care.
For investigational drug studies, changes in ADL performance
can be used as a secondary outcome measure to
document that cognitive or other effects of an anti-AD
Address correspondence and reprint requests to Dr. D. Galasko at
Neurology V127, VA Medical Center, 3350 La Jolla Village Drive, San
Diego CA 92161, U.S.A. (The Inventory of ADL and an accompanying
manual are available on request from Dr. Galasko.)
drug are clinically relevant. Treatment that enhances
cognitive function should lead to improvement in performance
of ADL, whereas treatment that slows the progression
or delays the onset of AD should be associated
with preservation or slower deterioration of ADL performance.
Although scores of ADL scales and cognitive tests are
correlated in patients with AD (Pfeffer et aI., 1982; Vitaliano
et aI., 1984), ADL performance also depends on
factors such as sustained attention, motivation, and motor
performance. It is difficult to predict from overall

S34 D. GALASKO ET AL.
cognitive test scores or even from tests of specific cognitive
domains which ADL are likely to be impaired or
how severe the impairment will be (Loewenstein et aI.,
1992). The amount of cognitive change needed to produce
a change in ADL performance is unknown.
ADL evaluation for clinical trials is an understudied
area. Most ADL scales were developed for general geriatric
assessment and to help determine the need for
services, and focus on basic activities such as walking,
feeding, and toileting (Katz et aI., 1963; Lawton et aI.,
1969). A few scales, such as those of Blessed (Blessed et
aI., 1968), Lawton (Lawton et aI., 1969), Pfeffer (Pfeffer
et aI., 1982), Weintraub (Weintraub, 1986), and the
NOSGER (Spiegel et aI., 1991) were introduced for geriatric
assessment or for clinical evaluation of patients
with AD, and include items that assess more complicated
activities (instrumental ADL, IADL). Some of these
scales contain gender-specific items or items performed
at infrequent intervals, such as filling out forms or documents.
Others ask whether a subject "can" perform an
ADL rather than relying on purely observed activities,
which introduces judgment or opinion into the informant's
report. Several scales combine both ADL and
behavioral information (Blessed et aI., 1968; Spiegel et
aI., 1991). Many IADL show a floor effect in AD, i.e.,
patients lose IADL very early in the course of dementia,
whereas basic ADL show a ceiling effect, with normal
performance until late in the course of AD (Spector et aI.,
1978). The approach of using loss of ADL or IADL as
milestones (Galasko et aI., 1995) is similar to that of the
Functional Assessment Scale (Reisberg, 1988) and is
best suited to clinical studies with long follow-up periods.
Performance ADL (PADL) scales require subjects to
perform ADL tasks in a structured setting using props
(Loewenstein et aI., 1992). Scoring is standardized and
can capture elements of performance such as sequencing,
initiation, and motivation. However, PADL scales cover
a relatively small number of activities, are timeconsuming,
and do not evaluate performance in the subjects'
own home environments. In view of the limitations
of existing instruments for use in clinical trials, the Alzheimer's
Disease Cooperative Study (ADCS) developed
a set of informant-based ADL.
METHODS
Development of the Inventory
A subcommittee of the ADCS, composed of clinicians
with expertise in dementia assessment and clinical trials,
developed an inventory comprising a wide range of
ADL. The initial item pool consisted of activities that
Alzheimer Disease and Associated Disorders, Vol. 11, Suppl. 2, 1997
normal elderly individuals regularly perform and that
were likely to be relevant to patients with AD over a
broad range of severity of dementia. The inventory included
ADL from existing scales and novel items based
on clinical experience, i.e., ADL necessary for personal
care, communicating and interacting with other people,
maintaining a household, conducting hobbies and interests,
and making judgments and decisions.
For each ADL, an informant is first asked whether or
not the patient attempted the activity during the past 4
weeks. If a patient did attempt the ADL, the informant is
asked to choose the single most accurate definition of the
patient's level of performance from a set of descriptions
of alternative methods of carrying out the ADL. For
ADL in which different methods of performance do not
apply, the informant is asked whether the subject usually
carried out the ADL "independently" (the highest
level), "with supervision" (needing verbal instructions
during ADL performance, an intermediate level of ability),
or "with physical help" (a lower level of performance).
If a subject needed reminders to get started but
then performed an ADL independently and successfully,
that is rated as independently. These definitions of levels
of independence are explained to informants before the
inventory is administered. To verify that nonperformance
of ADL was due to loss of ability, for each ADL that was
not attempted during the 4-week period, informants were
asked whether or not patients performed the activity
premorbidly.
The ADL questions were pretested at three sites,
modified where necessary to enhance their clarity,
and an inventory of 45 ADL was administered to AD
patients and controls as part of an ADCS multicenter
study.
Subjects
The ADCS study of instrument development for AD
clinical trials enrolled elderly controls (n = 64; 24 men
and 40 women) and subjects with AD (242 AD patients;
94 men and 148 women) as described (Ferris et aI., this
issue). Recruitment of subjects with AD was stratified by
Mini-Mental State Examination (MMSE; Folstein et aI.,
1975) scores at baseline, and all were communitydwelling
at the time of enrollment in the study. All subjects
were evaluated at baseline and at 6 and 12 months,
and half of the subjects were evaluated at 1 and 2
months. The present study made use of data at baseline,
1, 2, and 12 months.
Procedures
For each AD patient, an informant who spent at least
2 days per week with the patient was identified and in-

terviewed with the ADL Inventory. Controls acted as
their own informants. ADL interviews were conducted in
person at baseline. To increase follow-up participation, if
informants were unable to travel to a center's clinic site,
interviews were conducted by telephone. A representative
from each participating center attended a training
se~sion at which use of the inventory was described, and
ea~h site received a procedural manual describing each
question in the ADL Inventory. The personnel who carried
out ADL interviews were psychometrists or clinicians
with experience in assessment of patients with AD.
The ADL Inventory took from 30 to 45 min to administer.
Data Analysis
Data were analyzed using SPSS (version 6.1). We
used several criteria to determine which ADL in the inventory
were likely to be best suited for use in monitoring
patients in clinical trials. These included wide applicability,
good test-retest reliability, ability of the ADL to
show stepwise scaling and to correlate with dementia
severity, and decline of ADL performance in an appreciable
percentage of patients with AD over 12 months
(Feinstein et aI., 1986; Streiner and Norman, 1995). Applicability
was measured as the percentage of subjects
who attempted each ADL at baseline, or premorbidly in
the case of AD patients who did not attempt specific
ADL. As an index of short-term (test-retest) reliability
for each ADL, the extent of agreement between ratings at
baseline, one and two months was calculated using a K
statistic (O'Connell and Dobson, 1984). To determine
whether the levels of performing each ADL related to
dementia severity, we compared MMSE scores between
patients who carried out the ADL in each specified manner
at baseline by ANOVA. If MMSE scores of patients
performing an ADL at different descriptors or "levels"
of ADL performance did not differ significantly (p

S36 D. GALASKO ET AL.
TABLE 1. Metric properties of selected ADL Inventory questions in patients with AD
Applicability: % of subjects Change at 12 months:
attempting ADL at baseline
Test-retest Correlation
% of subjects who declined on ADL
ADL question
(number of levelst Overall
By MMSE stratab
>20 16 10 5 0
reliability
(K)
with MMSE
(Spearman R) Overall >20
By MMSE stratab
16 10 5 0
Handles mail (3) 30 1 1 I 0.41* 0.50** 22 .. .. 0 0
Discusses current events (4) 44 . 1 1 0.51* 0.48** 30 .. .. . 0
Makes a snack or meal (5) 50 . 1 1 I 0.62* 0.55** 30 0 .. .. 0
Writes note/letter/name (4) 51 . 1 1 I 0.59* 0.57** 33 0 .. .. .. 0
Uses household appliance (4)C 53 . 1 1 1 0.62* 0.61** 34 0 .. .. .. 0
Reads(4) 54 . 1 1 1 0.45* 0.56** 30 .. .. .. .. 0
Hobby/pastime/game (3r 56 . 1 1 1 I 0.59* 0.44** 29 0 0 .. .. 0
Chooses clothes to wear (3Y 56 . 1 1 I 0.59* 0.54** 36 0 .. .. .. 0
Turns off lights (2) 60 . 1 1 1 I 0.61* 0.44** 24 0 0 .. .. 0
Puts away belongings (4)C 61 . 1 1 1 I 0.45* 0.46** 35 0 .. .. .. 0
Can be left alone (4) 62 . 1 . 1 0.52* 0.55** 35 .. .. . .. 0
Goes shopping (3) 63 . 1 . 1 I 0.48* 0.47** 36 .. .. .. .. ..
Keeps appointments (4) 70 . 1 . 1 1 0.46* 0.44** 35 0 0 .. .. .
Obtainsa beverage(5) 70 . . . 1 I 0.54* 0.61** 40 .. . . .. 0
Finds belongings (3)C 71 . . 1 1 I 0.40* 0.42** 37 0 .. .. .. ..
Clears a table (4)C 71 . 1 . 1 1 0.60* 0.45** 32 0 0 .. .. ..
Disposes of litter (4)C 73 . . . 1 1 0.43* 0.49** 35 0 .. .. .. ..
Makes conversation (4) 75 . 1 1 1 1 0.41* 0.40** 35 0 0 .. .. ..
Grooms(5) 75 . . . 1 I 0.55* 0.64** 35 0 .. . 0
Gets dressed (5) 75 . . . 1 I 0.55* 0.66** 42 0 . . . 0
Bathes(5) 77 . . . 1 I 0.73* 0.66** 40 0 .. .. . ..
Uses a telephone (4) 80 . . . 1 1 0.50* 0.70** 47 .. . . . ..
Travelsoutsidehome(5) 90 . . . . . 0.59* 0.50** 44 .. 0 .. .. ..
Watches TV (5) 92 . . . . . 0.45* 0.44** 34 .. .. .. . ..
Toileting(4) 92 . . . . . 0.66* 0.61** 37 0 .. . .
Walking(5) 97 . . . . . 0.45* 0.28** 14 0 0 0
Eating(5) 97 . . . . . 0.60* 0.55** 30 0 .. . .
a ADL are arranged in ascending order according to the overall percentage of patients with AD who attempted the ADL at baseline.
b MMSE strata are as follows: >20, 20-28 points; 16,16-20; 10, 10-15; 5, 5-9; 0, 0-4. Symbols indicate the percentage of subjects in MMSE strata
who attempted each ADL at baseline: . >80%; 161-80%; 141-60%; 121-40%; blank 0-20%. Symbols for the percentage of subjects who declined
at 12 months are as follows: . 50%; .. 30-49%; 0 10-29%; blank 0-9%.
C ADL descriptorshad the format "independently,""with supervision,"or "with physicalhelp." All other ADL had specificallyworded
descriptors of different levels of ADL performance.
* p < 0.01; **p < 0.001.
baseline, a floor effect that renders it unsuitable for a
general-purpose ADL scale.
We next examined short-term (test-retest) reliability,
using K statistics to assess the extent to which ratings of
performance of each ADL item agreed at baseline and at
I and 2 months. Over 2 months, the progression of AD
should not be great enough to lead to a change in ADL
performance. In general, KSfell into the 0.4-0.75 range,
indicating moderate to very good agreement. Among
ADL with the highest KS were activities whose ratings
were dichotomous (yes or no) and the traditional basic
ADL. Several items, i.e., cleaning up spills, being aware
of toileting needs, and participating in a group event had
low KS «0.4, indicating only fair agreement) and were
eliminated from contention.
Stepwise scaling of levels of ADL performance and
correlation with dementia severity was the next criterion.
Tables I and 2 include the number of levels specified for
Alzheimer Disease and Associated Disorders, Vol. 11, Suppl. 2, 1997
each ADL in the inventory; an ADL question with a
dichotomous (yes or no) response has two levels, while
one that can be performed independently, with supervision,
with some physical help, or not at all has four. To
determine whether these levels have hierarchical properties,
we first examined the MMSE scores of patients who
performed each ADL at each specified level at baseline.
As an example, MMSE scores for levels of using a telephone
are displayed in Fig. 1. For most ADL, as each
level indicated less independent or less complex ADL
performance, the MMSE scores showed a pattern of progressive
decrease. If two descriptors of an ADL corresponded
to a similar range and distribution of MMSE
scores, then the descriptors were assumed to represent
cognitively equivalent levels of the ADL. This suggested
that the pair of descriptors could be merged and performance
of that ADL could be adequately described using
one level less. This situation arose very infrequently,

ADL INVENTORY S37
a ADL descriptors had the format "independently," "with supervision," or "with physical help." All other ADL had specifically worded
descriptors of different levels of ADL performance.
b Sev: impaired performance was noted only in severely demented patients.
almost exclusively for the descriptors that indicated
minimal performance of the ADL (e.g., "with help
only") and complete loss of the ADL.
To analyze scaling further, we calculated the Spearman
rank-order correlation coefficients between levels of
ADL performance and dementia severity, as indexed by
30
20
~
(IJ
::E
::E 10
TABLE 2. ADL with properties that did not meet specified criteria for assessment of ADL in clinical trials
Not attempted

S38 D. GALASKO ET AL.
patients with mild dementia (MMSE >20), to ensure that
change was adequately captured in that subgroup of patients.
Among the AD cohort overall, 12 ADL were relatively
insensitive to decline at 12 months, as listed in
Table 2. Although the rating of walking ability did not
meet this criterion, it is included in Table 1 because of its
traditional use as a fundamental basic ADL.
We next examined the relationship between baseline
MMSE severity stratum and the number of ADL items
(from the 27 in Table 1) on which subjects showed decline
of performance at 12 months (Fig. 2). Note that
subjects with AD in every MMSE stratum showed decline
on ADL, but subjects with moderate cognitive impairment
at baseline (MMSE 5-20) declined on more
ADL than did those with mild or very severe impairment.
Items Less Suited for ADL Assessment in a
Clinical Trial
Table 2 lists the ADL that failed to meet one or more
criteria. The most common problems with ADL in this
group were lack of wide applicability (usually due to a
gender effect), poor scaling, e.g., low correlation with
dementia severity, little change over 12 months, and
change restricted to severely demented patients only.
Most of these ADL had good reliability, and several of
the items in Table 2 would be suitable for use in customized
ADL scales for a clinical trial that targets
women, or for a study of severely impaired patients.
DISCUSSION
Many ADL items were widely applicable, had good
test-retest reliability, correlated significantly with the extent
of cognitive impairment in AD as measured by the
MMSE, and showed decline in performance in a significant
percentage of AD subjects at 12 months. These
properties indicate that they should prove valuable for
rating functional ability in AD clinical trials. The finding
that over one-third of the ADL failed to meet one or more
of these criteria illustrates the problems inherent in mea-
'" 0)
.S
U
0)
'"
25
20
-B
is
15
....
0....
10
0)
Z
5
0
Controls 21 - 30 16- 20 10- 15 5-9
Baseline MMSE group for subjects with AD
Alzheimer Disease and Associated Disorders, Vol. 11, Suppl. 2, 1997
0-4
suring ADL in a spectrum of patients with AD who are
typically the subject of clinical trials.
Most of the items showed good test-retest reliability
over 1-2 months, which presumably was aided by pretesting
the items to improve their clarity, conducting a
brief training session, distributing a procedures manual,
and specifying the amount of contact an informant had to
have with the patient. It is difficult to obtain consistent
ratings of intermediate levels of ADL performance, especially
when ADL have many potential levels or methods
of performance, even when these are specified as
carefully as possible. To improve test-retest reliability, it
may be necessary to improve the specific descriptors of
levels of performance, to collapse levels of performance
on some ADL, and to provide additional training to raters.
The validity of informant-derived descriptions of ADL
ability is difficult to establish directly. Observation of
subjects at home or in the community would be ideal but
is not practical. As a substitute, we estimated concurrent
validity of ADL reports by comparing ADL ability with
cognitive performance and found a substantial crosssectional
correlation between ADL performance and
MMSE scores for most items. Longitudinal decline on
many ADL items over 1 year among subjects with AD is
consistent with decline on cognitive and global measures
(Ferris et aI., 1997, this issue) and provides a further
indication of validity.
The optimal ADL performance by controls in this
study needs to be qualified. Among an elderly control
population, one would typically expect some degree of
ADL impairment as a result of physical disability, impaired
special senses, or medical illness. The controls in
this study were judged to be cognitively normal, and
represent high-functioning individuals living independently
in the community. Their results are therefore consistent
with expectations for a group of very healthy
elderly individuals but do not necessarily extrapolate to
the elderly in general, or to the "oldest old."
The large number of items in the ADL Inventory al-
FIG. 2. Number of ADL items in the Inventory (from
the 27 listed in Table 1) on which subjects with AD
showed decline in performance at 12 months. The box
plot shows the 25th, 50th (median), and 75th percentiles
as well as 95% limits for the total number of ADL on
which subjects in each MSE stratum declined.