1

The Health Assessment Questionnaire (HAQ)

and the Improved HAQ

(formerly called the PROMIS HAQ)

Stanford University School of Medicine
Division of Immunology & Rheumatology
(Revised June 2009)


This document describes the classic (Legacy) Health Assessment Questionnaire and
the Improved HAQ (formerly called the PROMIS HAQ), their history, administration and
scoring conventions, and provides information regarding user permission.

The Heath Assessment Questionnaire (HAQ)
The HAQ was originally developed three decades ago by James F. Fries, MD, and
colleagues at Stanford University. Conceived to represent a model of patient-oriented
outcome assessment, the HAQ is one of the first instruments to measure patient
reported outcomes (PRO).
The HAQ is widely used throughout the world. It has been administered in rheumatic
disease studies, HIV/AIDS patients, normal aging cohorts, and in population-based
studies, including the follow-up to the National Health and Nutrition Examination Survey
(NHANES). We refer you to reviews in 1992, 1996, and 2003 (see references), which
examine the HAQs history and document its reliability, validity, and applicability in
multiple settings and languages.
The HAQ has been validated in patients with a wide variety of rheumatic diseases,
including rheumatoid arthritis, osteoarthritis, juvenile rheumatoid arthritis, lupus,
scleroderma, ankylosing spondylitis, fibromyalgia, and psoriatic arthritis. However, the
HAQ has been and can be administered in diverse disciplines and in different cultures
without any impact on its reliability or validity with properly designed adaptations.
The HAQ has played a major role in numerous areas, such as in the prediction of
successful aging, inversion of the therapeutic pyramid in rheumatoid arthritis (RA),
2
quantification of NSAID gastropathy, development of risk factor models for
osteoarthrosis, and examination of mortality risks in RA.
The HAQ has become a de facto mandated outcome measure for clinical trials in
rheumatoid arthritis and some other diseases. Although it has sometimes been
characterized as a disease-specific instrument rather than having been adjudicated on
the basis of its structure, content, and history, the HAQ should be considered generic
rather than disease-specific.
THE HAQS TWO VERSIONS - THE 2-PAGE (SHORT) HAQ-DI AND THE FULL HAQ
The HAQ has been administered primarily in one of two versions, the 2-page or short
HAQ-DI (Disability Index) or the Full HAQ. The 2-page HAQ-DI has received the widest
attention, the most frequent use, and is the most cited version. It assesses the extent of
a patients functional ability. It is commonly called the HAQ, and consists of the HAQ-
DI, along with the HAQ Pain Visual Analog Scale (VAS), and the Patient Global Health
VAS. These are core items that have not changed since the HAQs inception.
The Full HAQ is a multi-dimensional instrument. It contains five sections which address
generic health dimensions, or the 5 Ds: (1) disability; (2) discomfort; (3) drug toxicity;
(4) health care utilization and cost (dollars); and (5) death. It also includes
supplemental items on demographics, lifestyle, and health behaviors.
Other than the HAQ-DI and the pain and patient global health scales, the additional
items on the full HAQ are used primarily by the Arthritis, Rheumatism, and Aging
Medical Information System (ARAMIS) to address research areas of interest. Items are
tailored and modified as required to address specific hypotheses or research questions.
RELIABILITY AND VALIDITY OF THE HAQ-DI
The HAQ-DI is sensitive to change and is a good predictor of future disability and costs.
It has demonstrated reliability and validity in different languages and contexts. Test-
retest correlations range from 0.87 to 0.99. Correlations between interview and
questionnaire format range from 0.85 to 0.95. Validity has been demonstrated in
literally hundreds of studies.
There is consensus that the HAQ-DI possesses face and content validity. Correlations
between questionnaire or interview scores and task performance range from 0.71 to
0.95 demonstrating criterion validity. The construct/convergent validity, predictive
validity, and sensitivity to change have also been established in numerous observational
studies and clinical trials. The HAQ-DI has also demonstrated a high level of convergent
validity based on the pattern of correlations with other clinical and laboratory measures.
3
The HAQ-DI is designed to assess the patients USUAL abilities using their usual
equipment over the past week. Some patients have questioned whether their responses
should reflect a particularly good or bad time, which may be out of the time frame
requested, because they feel that their response may be missing those times when their
functional ability changes. However, by repeating the HAQ at specific and regular time
intervals (ARAMIS administers the HAQ every 6 months), patterns of function can be
examined. Inquiring about these activities only when patients are feeling particularly
good or bad would result in inaccurate and biased data.

ADMINISTRATION
The Full and the 2-page HAQ are usually self-administered. However, either one can be
administered face-to-face in a clinical setting or in a telephone interview by a trained
interviewer, Both have been validated in these settings.
In ARAMIS, the questionnaire is mailed to patients every six months. They are asked to
complete it without additional instructions.
The Full HAQ takes 20 to 30 minutes to complete. The 2-page HAQ or the Improved HAQ
can be completed in approximately five minutes. Patients usually find the instruments
self-explanatory. Clarification is seldom required.

STRUCTURE HAQ-DI
The HAQ-DI is composed of 20 items in 8 categories (Dressing and Grooming, Hygiene,
Arising, Reach, Eating, Grip, Walking, Common Daily Activities [IADL} also see Figure at
end of this document). Items were developed to be mutually exclusive and collectively
exhaustive of activities related to physical functioning. Each category has at least two
sub-category questions. Within each category, patients report the amount of difficulty
they have in performing the specific sub-category items. There are four response
options ranging from No Difficulty to Unable to Do, scored 0-3.

HANDLING RESPONSES
Patients should respond to items using their own frame of reference. Thus, the ratings,
No, Some, Much, or Unable, are deliberately not operationalized. For example, if a
4
patient asks what Some means, an appropriate response would be Whatever you
think Some means to you.
When a sub-category item does not apply, e.g., when a patient does not shampoo their
hair (perhaps they have no hair), they should give the response that would represent
the ability they believe they would have had they been able to attempt the activity.
When a patient uses aids or devices (e.g., crutches, jar openers, etc.), then they answer
the question based on their usual equipment or way of performing the activity. If they
have no difficulty with a sub-category item when using aids/devices (see below for
discussion of aids/devices), then they mark the no difficulty column. In scoring, the
use of aids/devices results in an adjusted score for that item (see below).
Improperly Marked Items: When a patient makes a mark between two response
options, then the score is the closest one. When it is directly between two response
options, use the higher number.
Items Left Blank: When all sub-category items in a category are left blank, or when more
than one response is given, then follow up with the patient is required.
Note: A HAQ-DI score cannot be calculated validly when there are scores for less than
six of the eight categories.

SCORING
The HAQ-DIs scoring conventions allow for the computation of two disability indices:
1. The Standard HAQ-DI, the preferred and traditional scoring method, which takes
into account the use of aids/devices.
2. The Alternative Disability Index, which does not.
In any event, a patient must have a score for at least six of the eight categories.
Otherwise a HAQ-DI score cannot be validly computed.
Computing the Standard HAQ-DI Score (With Aids/Devices): There are three steps:
1. Sum the 8 category scores by using the highest sub-category score from each
category.
5
For example, in the category ARISING there are three sub-category items. A
patient responds with a 1, 2, and 0, respectively; the category score is 2.
2. Adjust for use of aids/devices and/or help from another person when indicated.
The Table below identifies the Aid/Devices companion variable for each HAQ-DI
category.
Adjust the score for a category by increasing a zero or a one to a two.
If a patient's highest score for that sub-category is a two it remains a two, and
if a three, it remains a three.
3. Divide the summed category scores by the number of categories answered (must
be a minimum of 6) to obtain a HAQ-DI score of 0-3 (3=worst functioning).

Companion Aids/Devices Items for HAQ-DI Categories
HAQ-DI Category Companion Item

Dressing & Grooming

Devices used for dressing (button hook, zipper
pull, long handled shoe horn etc.)
Arising Built up or special chair
Eating Built up or special utensils
Walking Cane walker, crutches
Hygiene Raised toilet seat, bathtub seat, bathtub bar
Long handled appliances in bathroom
Reach Long handled appliances for reach
Grip Jar opener (for jars previously opened


Notes about Aids/Devices:
The assignment of devices to particular disability categories assumes that the devices
are used only for their intended purposes, e.g., when a patient indicates that they
use a cane, it is presumed that the cane is used as an aid in walking. However, it is
possible that patient uses the cane as an aid in performing other activities. For
example, the patient may check off the cane listed at the bottom of the page (or
write cane under the other option) and then write a little note in the margin
stating that the cane is also used on a regular basis to help them rise out of a chair
and to rise off of the toilet. In such a case, the variables should be scored a 1 to
reflect the patients use of a cane in these three areas of functioning. If unsure
whether the patient is using one of the devices specified above for the purpose for
which it is designed, the patient should be called for clarification.
6
When there are devices entered in the Other section or notes written next to an
item, they are considered as being used for any of the stated categories. Permanent
adaptations of the persons environment (e.g., changing faucets in the bathroom or
kitchen, using fabric hook and loop closures on clothing) should also be counted as
aids/devices.
In ARAMIS, few patients have reported other items, and when they have, it has
usually been either a duplicate of an aid/device already listed, or they have listed
something that does not count (e.g., a wrist splint).Thus, it is usually acceptable to
exclude the Other option, if desired.
Computing the Alternative HAQ-DI Score (Without Aids/Devices)
The Standard scoring method is preferred and the most common. However, there are
circumstances where the alternative score would be preferable, such as when an
investigators goal is to increase the use of aids/devices. If aids/devices were taken into
account, scoring would be biased, and findings would indicate that the study increased
HAQ disability scores.
There are two steps in computing the alternative HAQ-DI:
1. Sum the scores for each category, ignoring scoring for aids/devices.
2. Divide by the number of categories answered (minimum of 6 required). This
yields a score of 0-3 (3=worst disability).

7


The Improved HAQ

(formerly called the PROMIS HAQ)


The Improved HAQ (formerly called the PROMIS HAQ) is the first substantial revision of
the HAQ-DI since its inception three decades ago. We have revised its name to avoid
confusion with official instruments of the Patient Reported Outcomes Measurement
Information System (PROMIS). Please see the PROMIS website
(http://www.nihpromis.org) for information about the PROMIS Physical Function item
bank and short forms and the PROMIS Assessment Center
sm
.

Background
The Improved HAQ evolved from its parent, the original Health Assessment
Questionnaire Disability Index (HAQ-DI).
In the mid-2000s, in collaboration with QualityMetric, Inc. we became one of the six
primary research sites in the Patient Reported Outcome Measurement Information
System (PROMIS) (www.nihpromis.org ). PROMIS is a component of the National
Institutes of Health (NIH) Roadmap Initiative aimed at improving PROs and promoting
their broader use by the private and public sectors. Its goal is to improve the quality and
efficiency of clinical research; in part by reducing sample sizes required in clinical trials.
The Improved HAQ came about in conjunction with our work with PROMIS and followed
standardized PROMIS protocols for item development. Original HAQ-DI items
underwent qualitative and quantitative psychometric processes to improve their
measurement properties. Qualitative evaluation involved a stepwise process that
included expert review, patient surveys, focus groups, and cognitive interviewing. Items
were then rewritten, field tested, and quantitatively assessed using advanced
psychometric methodology.
Findings from these analyses indicated that an item was clearer when written in the
present tense without a time frame and with more response options. We rewrote
original HAQ-DI items by deleting the one-week time frame, retaining the present tense
8
capability framework (Are you able to ...),and increasing the response option set from
the original four choices (Without any difficulty; with some difficulty; with much
difficulty; and unable to do) to five by adding with a little difficulty. Results of
analyses showed that the addition of the fifth response option decreased the number of
responses to without any difficulty, thereby reducing ceiling effects. This process
resulted in HAQ-DI items with better structure, clarity, and reliability.
The Improved HAQ contains the same 20 items as the original HAQ-DI. The Figure at the
end of this document gives a visual comparison of the two instruments. Overall, its
items are clearer, simpler, more clearly understood, assessed more quickly, have less
floor and ceiling effects, have greater information content, smaller measurement errors,
and are more reliable with better face and content validity and better responsiveness
than their original HAQ-DI counterparts. We also reduced aids/devices from 21 to 4 and
found that there is no significant effect on scores. We have initially set the scoring scale
at 0-100 for ease in computation. However, we continue to investigate alternative
scoring algorithms.
The figure below shows the comparison of information content between an original
HAQ-DI item and an Improved HAQ item (black triangles), which contains five instead of
four response options. The scale has zero set at the mean of a normal population, and
each integer represents one standard deviation from that mean. The height of the curve
above the standard error of zero represents the item information content. The
information content of the Improved HAQ item (black triangles) is spread more broadly,
meaning that it provides information across a greater range of function. The data show
relevant improvements for the Improved HAQ item with respect to measurement range
and peak information.
0.00
2.00
4.00
6.00
8.00
-4.00 -3.00 -2.00 -1.00 0.00 1.00 2.00
Theta
I
t
e
m

i
n
f
o
r
m
a
t
i
o
n

9
We recently compared internet with mailed administrations of the original HAQ-DI with
the Improved HAQ. We studied response rates, instrument scores, data validity, and
sensitivity to change at one year in a group of RA and osteoarthritis participants. We
found that the HAQ-DI and the Improved HAQ were similar relative to completion rates
and instrument scores, and that they measured the same constructs. Both instruments
were sensitive to change, although the Improved HAQ was more responsive than the
HAQ-DI after one year. (manuscript under review).
Use of the Improved HAQ in clinical research is encouraged, with the understanding that
data collected from that use will contribute further to knowledge about the validity of
these measures.

STRUCTURE Improved HAQ

The Improved HAQ is comprised of the same contextual 20 items contained in the HAQ-
DI. All items are better than their original counterparts in all attributes; hence, the
Improved HAQ is more reliable, valid, and sensitive to change than the HAQ-DI. We
recommend its use in all instances where the HAQ-DI would have been used except in
studies in progress where baseline measurements used the HAQ-DI.
Modifications in wording from the HAQ-DI were made from findings in item analyses.
The items do not contain a time frame, are written in the present tense, and have five
response options (the HAQ-DI contains four). Patients respond from 0-4, with 4=worst
functioning.
In contrast to the HAQ-DI, items in the Improved HAQ are not grouped by physical
function category. Rather than listings of aids/devices or assistance as in the HAQ-DI,
the Improved HAQ uses four questions asking about use of aids/devices or assistance.
The Improved HAQ instrument also includes the HAQ pain and patient global health
scales. They are handled as described below.

ADMINISTRATION
As with the HAQ-DI, patients should use their own frame of reference when responding
to items. Thus, the terms None, A Little, Some, Much, or Unable, are deliberately not
operationalized. For example, when a patient asks what Some means, an appropriate
10
response would be Whatever Some means to you. When an item does not apply to
a patient, e.g., they dont shampoo their hair (perhaps they have no hair), they give the
response that would represent the ability they believe they would have had they
attempted the activity.
HANDLING RESPONSES
Improperly Marked Items. When a patients mark is between the response columns,
then the score is the closest one. If its directly between, use the higher number.
Items Left Blank. When more than one response to an item is given, then follow-up with
the patient is required.
Note: An Improved HAQ score cannot be calculated validly if there are responses to
fewer than 15 of the 20 items.

SCORING

The Improved HAQ results in a score from 0-100 and is computed as follows:
1. Each item is scored a 0, 1, 2, 3, or 4.
2. If any of the aids/devices items are answered yes, then first adjust the score. If
it is a 0, 1, or 2 for any of the associated items (see below for listing) the score for
that item is raised to a 3 before computing a score. An item scored 4 remains a 4.
3. Sum the 20 items and divide by 20 to yield an average score from 0 to 4. Multiply
by 25 to convert the score to 0-100.
Alternative Scoring Methods
Two alternative scoring methods are being evaluated in parallel and offer some
advantages. Ultimately, we will recommend the method that is most sensitive to
change.
1. Using all 20 items, divided into 8 categories. After #1 above average the highest
items in each of the 8 categories, then multiply by 25.
2. Using 16 items, 8 categories.
11
The following items are deleted from the 20-item Improved HAQ to comprise a 16-item
HAQ:
(1) stand up from an armless straight chair
(2) cut your food using eating utensils
(3) walk a block on flat ground
(4) reach and get down a 5-pound object from above your head.

After completing step 1 above, average the highest items in each of the 8 categories,
then multiply by 25.
Aids/Devices
The table below shows the Improved HAQs four aids/devices items with their
companion Improved HAQ items and associated HAQ-DI categories. There are no
associated aids/devices question for Improved HAQ items #2, 4, 5 -7, 10, 11, 14-17 ,19,
or 20 (See HAQ instrument).



Improved HAQ Aids/Devices Items
Companion
Improved
HAQ Items
Associated
HAQ-DI
Category(ies)

1. To get around, do you usually need a cane, crutches,
walker, wheelchair, or help from another person?


8, 9, 1

Walking
2. To stand up from a sitting position, do you usually
need a special chair, a raised toilet seat, or help from
another person?

3,12

Arising
Hygiene
3. To get dressed, do you usually need a buttonhook,
zipper pull or other gadget, or help from another
person?

1 Dressing/
Grooming
4. To reach something, do you usually use long-handled
appliances or help from another person?
13 Reach
Hygiene


12

USER PERMISSION

The HAQ-DI and the Improved HAQ are copyrighted to insure that they will be used
unmodified to preserve the validity of results and contribute to standardization of
assessment across studies. However, both are considered to be in the public domain,
with the request that users cite relevant articles(s) in their publications (see references
at the end of this document).
There is no charge from Stanford for permission to use the HAQ-DI or the Improved
HAQ. However, other groups which have independently translated the HAQ may charge
for their use.
Permission from PROMIS is not needed to use the Improved HAQ.
13

FIGURE. Content and Configuration Comparison of the HAQ-DI and The Improved HAQ
Items common to the HAQ-DI and the Improved HAQ
1. Dress yourself, including shoelaces and buttons?
2. Shampoo your hair?
3. Stand up from a straight chair
4. Get in and out of bed
5. Cut your own meat?
6. Lift a full cup or glass to your mouth?
7. Open a new milk carton?
8. Walk outdoors on flat ground?
9. Climb up five steps?
10. Wash and dry your body?
11. Take a tub bath?
12. Get on and off the toilet?
13. Reach and get down a 5 pound object (such as
a bag of sugar) from above your head?
14. Bend down to pick up clothing from the floor?
15. Open car doors?
16. Open previously opened jars
17. Turn faucets on and off?
18. Run errands and shop?
19. Get in and out of a car?
20. Do chores such as vacuuming or yard work?
Instrument HAQ-DI Improved HAQ
Time Frame (past tense) Over the past week, are you
able to
(present tense) Are you able to
Response
Options
(n=4)
Without any difficulty, With some
difficulty, With much difficulty, Unable to
do
(n=5)
Without any difficulty, With a little difficulty,
With some difficulty, With much difficulty,
Unable to do
Aids/Devices (n=21) Please check any categories for
which you usually need help from
another person:
1. Dressing and grooming
2. Eating
3. Walking
4. Arising
Please check any aids or devices that
you usually use:
5. Devices used for dressing (button
hook, zipper pull, etc.),
6. Built up or special utensils
7. Crutches
8. Cane
9. Walker
10. Wheelchair
11. Special or built up chair
12. Raised toilet seat
13. Long-handled appliances for reach
14. Bathtub seat
15. Bathtub bar
16. Long-handled appliances in
bathroom
17. Jar opener (for jars previously
opened)
Please check any categories for which
you usually need help from another
person:
18. Hygiene
19. Reach
20. Gripping and opening things
21. Errands and chores
(n=4) For each of the following questions,
please answer by placing an X in the
appropriate box: [Response options - Yes/No]
1. To get around, do you usually need a
cane, crutches, walker, wheelchair, or
help from another person?
2. To stand up from a sitting position, do you
usually need a special chair, a raised toilet
seat, or help from another person?
3. To get dressed, do you usually need a
buttonhook, zipper pull or other gadget, or
help from another person?
4. To reach something, do you usually use
long-handled appliances or help from
another person?



14
The HAQ Pain Visual Analog Scale (VAS)

The he HAQ Pain Scale assesses arthritis-related pain and its severity over the PAST
WEEK on a double-anchored VAS (a horizontal line where each end represents opposite
ends of a continuum). It obtains data on how arthritis-related pain has USUALLY been
over the past week, even though pain may be reported to vary over the course of a day
or from day to day.

The VAS line is standardized to 15 centimeters in length, which is convenient for the
page and the patient. The scale is labeled from zero (no pain) at the left anchor point
and 100 (severe pain) at the right anchor point.

Patients are instructed to place a vertical mark on the line to indicate the severity of
their pain.

HANDLING RESPONSES

When a patient writes in a number on the pain scale, or writes a number in
addition to making a mark, use the written number, converting it to the
corresponding score. In this case, do not measure the mark. For example, if the
patient writes 50 on the line, this should be coded as 1.5 (see below for scoring
computation).

When a patient records a percentage, multiply the percentage by 3.

When a patient puts more than one mark, the midpoint is used.

When a patient makes a horizontal line below the pain scale, instead of a vertical
one, the midpoint of that line is taken. If the line starts at the beginning of the
scale, measure to the end of the line not the middle.

COMPUTING THE PAIN SCORE

Using a metric ruler, from zero to the patients mark, measure the distance in
centimeters and multiply by 0.2. This converts centimeters into the appropriate metric
and yields a score from 0 to 3. For example, if the mark is at 8 centimeters, 8 x 0.2 yields
a pain score of 1.6. The table below shows measurement conversions.

Alternatively, a 0-100 mm scale may be used, which requires no score conversion.
15

Conversion of Centimeter Measurements to Pain Scores
Centimeters Score Centimeters Score Centimeters Score
0 = 0 4.8 - 5.2 = 1.0 10.3 - 10.7 = 2.1
0.1 - 0 .7 = 0.1 5.3 - 5.7 = 1.1 10.8 - 11.2 = 2.2
0.8 - 1.2 = 0.2 5.8 - 6.2 = 1.2 11.3 - 11.7 = 2.3
1.3 - 1.7 = 0.3 6.3 - 6.7 = 1.3 11.8 - 12.2 = 2.4
1.8 - 2.2 = 0.4 6.8 - 7.2 = 1.4 12.3 - 12.7 = 2.5
2.3 - 2.7 = 0.5 7.8 - 8.2 = 1.6 12.8 - 13.2 = 2.6
2.8 - 3.2 = 0.6 8.3 - 8.7 = 1.7 13.3 - 13.7 = 2.7
3.3 - 3.7 = 0.7 8.8 - 9.2 = 1.8 13.8 - 14.2 = 2.8
3.8 - 4.2 = 0 .8 9.3 - 9.7 = 1.9 14.3 - 14.7 = 2.9
4.3 - 4.7 = 0 .9 9.8 - 10.2 = 2.0 14.8 - 15.0 = 3.0



THE HAQ PATIENT GLOBAL VAS

The HAQ Patient Global Health VAS is a validated measure of quality of life. Fries and
Ramey compared the HAQ Global to the Torrance quality-of-life feeling thermometer
and found the two scales to be highly correlated (r = -0.676; p<0.001), indicating that
both instruments were measuring similar quality of life constructs (see Fries and Ramey
1997).

The Global VAS is also a 15-centimeter, double-anchored horizontal scale that starts at 0
(very well) and goes to 100 (very poor).

Handling responses and scoring are the same as the Pain Scale (please see above).

16

References

Bruce B, Fries JF The Stanford health assessment questionnaire (HAQ): a review of its
history, issues, progress, and documentation. J Rheumatol 2003;30(1): 167-178.
Cella D, Yount S, Rothrock N, Gershon R, Cook K, Reeve B, et al. The patient reported
outcomes measurement information system (PROMIS): Progress of an NIH roadmap
cooperative group during its first two years. Medical Care 2007;45:S3-S11.
Fries JF, Bruce B, Cella D. The promise of PROMIS: using item response theory to
improve assessment of patient-reported outcomes. Clin Exp Rheumatol 2005;23(5 Suppl
39):S53-7.
Fries JF, Bruce B, Cella D. The promise of PROMIS: using item response theory to
improve assessment of patient-reported outcomes. Clin Exp Rheumatol 2005;23(5 Suppl
39):S53-7.
Fries JF, Spitz PW, et al. Measurement of patient outcome in arthritis. Arthritis Rheum
1980;23(2): 137-45.
Fries JF, Spitz PW, et al. The dimensions of health outcomes: the health assessment
questionnaire, disability and pain scales. J Rheumatol 1982;9(5): 789-93.
Fries, JF, Ramey DR. "Arthritis specific" global health analog scales assess "generic"
health related quality-of-life in patients with rheumatoid arthritis. J Rheumatol
1997;24(9): 1697-702.
Ramey D, Fries JF, et al. (1995). The Health Assessment Questionnaire 1995 - status and
review. quality of life and pharmacoeconomics in clinical trials. S. B. Philadelphia,
Lippincott-Raven Publishers: 227-237.
Ramey DR, Raynauld JP, et al. The health assessment questionnaire 1992: status and
review" Arthritis Care Res 1992;5(3): 119-29.