/clinical/,/clinical/cckm-tools/,/clinical/cckm-tools/content/,/clinical/cckm-tools/content/questionnaires/,/clinical/cckm-tools/content/questionnaires/related/,

/clinical/cckm-tools/content/questionnaires/related/name-101352-en.cckm

201606180

page

100

UWHC,UWMF,

Clinical Hub,UW Health Clinical Tool Search,UW Health Clinical Tool Search,Questionnaires,Related

Development and testing of the MIDAS Questionnaire to assess headache-related disability

Development and testing of the MIDAS Questionnaire to assess headache-related disability - Clinical Hub, UW Health Clinical Tool Search, UW Health Clinical Tool Search, Questionnaires, Related


Development and testing of the Migraine
Disability Assessment (MIDAS)
Questionnaire to assess headache-related
disability
Walter F. Stewart, MPH, PhD; Richard B. Lipton, MD; Andrew J. Dowson, MB; and James Sawyer, MB
Article abstract—The MIDAS Questionnaire was developed to assess headache-related disability with the aim of
improving migraine care. Headache sufferers answer five questions, scoring the number of days, in the past 3 months, of
activity limitations due to migraine. The internal consistency, test–retest reliability, and validity (accuracy) of the
questionnaire were assessed in separate population-based studies of migraine sufferers. In addition, the face validity, ease
of use, and clinical utility of the questionnaire were evaluated in a group of 49 physicians who independently rated disease
severity and need for care in a diverse sample of migraine case histories. The test–retest Pearson correlation coefficient for
the total MIDAS score was approximately 0.8. The MIDAS score was valid when compared with a reference diary-based
measure of disability; the overall correlation between MIDAS and the diary-based measure was 0.63. The MIDAS score
was also correlated with physicians’ assessments of need for medical care (r 5 0.69). From studies completed to date, the
MIDAS Questionnaire has been shown to be internally consistent, highly reliable, valid, and correlates with physicians’
clinical judgment. These features support its suitability for use in clinical practice. Use of the MIDAS Questionnaire may
improve physician–patient communication about headache-related disability and may favorably influence health-care
delivery for migraine patients.
NEUROLOGY 2001;56(Suppl 1):S20–S28
Measuring migraine severity helps target patients
who differ in their treatment needs.
1
However, mea-
suring severity is challenging because of the episodic
occurrence of attacks and variation in impact from
one attack to the next.
2
In a single individual, head-
aches may vary in level of pain, duration, associated
symptoms, and disability.
Migraine is not always managed effectively by
physicians and continues to be under-diagnosed and
under-treated in clinical practice.
3
Physicians and
patients often do not communicate effectively about
functional limitations imposed by a range of ill-
nesses, including migraine. An instrument that as-
sesses migraine severity has the potential to improve
migraine care by facilitating physician–patient com-
munication and guiding treatment decisions.
To be useful for clinical practice, a migraine sever-
ity instrument should, at minimum, meet a number
of key research and clinical practice criteria (table
1).
3
Briefly, the research criteria require the instru-
ment to be reliable, valid, and internally consistent.
If it is to be used to follow patients, it must be sensi-
tive to change (table 1). For use in clinical practice,
the tool also must be easy to use and score and
should be scored in intuitively meaningful units for
patients and physicians (table 1). Two related ap-
proaches to assess the severity of migraine have
been used: measuring health-related quality of life
(HRQoL) and measuring headache-related disability.
HRQoL instruments measure global aspects of an
individual’s health status over a period of time, in-
cluding time with and without illness.
4
HRQoL tools
may be generic or disease-specific.
5-9
Generic tools,
such as the SF-36, are useful for a broad range of
specific illnesses. They measure a range of domains
and are particularly useful for comparing the burden
posed by various illnesses. In contrast, disease-
specific HRQoL instruments focus on the specific
burdens of particular illnesses. These tend to be
more sensitive to change and highly relevant to long-
term treatment trials. Disease-specific HRQoL in-
struments tested in migraine exhibit good reliability
and show evidence of sensitivity to change.
10
How-
ever, HRQoL questionnaires are often lengthy, com-
plicated to score, and are expressed in arbitrary
units. As a consequence, the HRQoL questionnaires
are valuable research tools but are not optimal for
routine severity assessment in clinical practice, for
From the Johns Hopkins School of Public Health, Baltimore, MD and Innovative Medical Research, Inc., Hunt Valley, MD (Dr. Stewart); the Albert Einstein
College of Medicine, Bronx, NY and Innovative Medical Research, Inc., Stamford, CT (Dr. Lipton); the Kings Headache Service, Kings College Hospital,
London, UK (Dr. Dowson); and Willaston, Cheshire, UK (Dr. Sawyer).
AstraZeneca provided W.F.S., R.B.L., A.J.D., and J. Smith an honorarium for their participation in this project. AstraZeneca has provided grant support in
excess of $10,000 to W.F.S. and R.B.L.
Address correspondence and reprint requests to Dr. Walter Stewart, Innovative Medical Research, Inc., 1200 High Ridge Road, Stamford, CT 06905. e-mail:
bstewart@imrinc.com
S20 Copyright © 2001 by AAN Enterprises, Inc.

which brevity, simplicity and ease of interpretation
are essential.
In contrast to HRQoL measures, disability is a
considerably less complex and more accessible con-
cept. Disability is a consequence of illness and an
important indicator of unmet treatment need. In
part, disability is the product of poor quality of life
that is attributable to illness. The World Health Or-
ganization (WHO) defines disability in relation to
work and to function in other roles.
11
In 1993, the
WHO launched an initiative on neurology and public
health that included measuring the global burden of
disease using Disability Adjusted Life Years
(DALYS).
12
The DALY is measured in units of lost
time due to a specific illness. The WHO also engaged
in an assessment of disease severity using the person
trade-off method. They rated 22 “indication condi-
tions” from 0.00 (perfect health) to 1.00 (equivalent
to death) and classified the conditions into seven cat-
egories. The most disabling group of conditions in-
cluded severe migraine and three other disorders
(dementia, quadriplegia, and acute psychosis). Se-
vere migraine was considered more disabling than
several other conditions, including blindness, para-
plegia, angina, and rheumatoid arthritis.
13-15
The burden of headache disorders is well docu-
mented. Migraine can result in missed work and
household work, reduced productivity in work roles,
and lost time in non-work activities.
16-19
The degree
of disability is related to both the direct and indirect
costs of illness.
20
In addition, information on disabil-
ity is an important determinant of physicians’ judg-
ments of migraine severity and treatment needs.
21,22
Lipton, Stewart, Von Korff et al. have focused on dis-
ability in formulating tools to assess migraine severity.
A series of related disability instruments have been
developed, including the Chronic Pain Index (CPI),
23
the Headache Impact Questionnaire (HlmQ),
16,24
and
the Migraine Disability Assessment (MIDAS)
Questionnaire.
25-27
The CPI measures severity of a number of pain
disorders
23
by combining measures of pain and dis-
ability. The CPI score exhibited excellent prognostic
validity in a prospective study of managed care pa-
tients with a variety of pain conditions (e.g., back
pain, headache pain).
23
However, the CPI scoring is
complex and the combined pain–disability scale
lacks intuitive meaning.
The HImQ was specifically designed to assess
headaches. Similar to the CPI, the HImQ combined
measures of pain intensity and disability, together
with headache frequency. Population-based studies
showed that the HImQ was highly reliable and valid
for a 3-month recall period.
16,24
However, the calcula-
tion formula for the HImQ score was too complicated
for self-scoring and for use in primary care because it
involved several steps including multiplication and
addition. In addition, because the score was based on
a composite of different measures, it could not be
expressed in intuitively meaningful units, a factor
that limits its practical utility.
The MIDAS Questionnaire was specifically devel-
oped to improve on the practical limitations of the
HImQ.
Development of the MIDAS Questionnaire.
The MIDAS Questionnaire measures headache-
related disability based on five disability questions
(figure 1). Patients record the number of missed days
due to headache in school or paid work (Question 1),
household work (Question 3), and family, social, or
leisure activities (Question 5). These domains were
selected because of their importance to 20- to 50-
year-olds, the age group of highest prevalence for
migraine.
28
Two further questions assess the number
of additional days with significant limitations to ac-
tivity (defined as at least 50% reduced productivity)
in the domains of employment (Question 2) and
household work (Question 4). The latter question for-
mat was designed to capture days with substantial
Table 1 Research and clinical practice criteria for evaluating
clinical measures for headache severity
3
Criteria Analysis
Research criteria
Internal consistency The extent to which the items
comprising the measure are related
to each other
Test–retest reliability Stability and reproducibility of results
when the instrument is
administered twice to the same
person (test–retest)
Content validity Correlation between the instrument-
based measure and a gold standard
measure
Construct validity The extent to which an instrument
measures what it purports to
measure and fits into a theoretical
scheme about the variable of
interest
Discriminant validity The extent to which an instrument
distinguishes two conditions or
states known to be different
External validity The extent to which the instrument-
based measure is related to other
measures considered to be relevant
Sensitivity to change Instrument detects real change over
time, such as improvement in
outcome in response to effective
therapy
Clinical practice criteria
Face validity Judgment that the measure
corresponds to an individual’s
perception, e.g., by selecting items
deemed to be important to the
disease sufferer or physician
Ease of use The instrument should be simple to
use, score, and interpret
Intuitively
meaningful
The instrument should correlate with
physicians’ judgments of illness
severity and treatment need
March 2001 NEUROLOGY 56(Suppl 1) S21

reduction in productivity. Days with productivity re-
duced by half or more are given a full count to com-
pensate, in part, for not counting days in which
productivity is reduced by less than half. This for-
mat, however, was not considered useful in assessing
lost time for family, social, and leisure activities be-
cause it was difficult for subjects to interpret.
The MIDAS score is derived as the sum of lost
days due to headache recorded for Questions 1 to 5.
The MIDAS score can sometimes be higher than the
actual number of lost headache days if more than
one domain of activity is affected on the same day.
Under these circumstances, the score reflects the
range of demands in an individual’s life whether or
not an individual works for pay. Although individu-
als who do not work for pay answer only three ques-
tions, the two relevant domains are likely to be more
important than they are for those who do not work
for pay. In previous work, the HImQ score for indi-
viduals who worked for pay did not differ from those
who did not work for pay.
16,24
Two additional MIDAS questions (A and B) collect
information on the frequency of headaches and the
intensity of the headache pain. These are not scored
in the MIDAS Questionnaire but are included to pro-
vide the physician with clinically relevant informa-
tion that may aid in treatment and management
decisions.
A 3-month recall interval was selected for all
MIDAS questions to balance the accuracy of self-
reported information (optimized by a short recall in-
terval) with the representativeness of the reported
headache experience over time (optimized by a long
recall interval). Previous work has shown that 3
months was a short enough period for accurate re-
call.
24,27
Testing the MIDAS Questionnaire. The MIDAS
Questionnaire was evaluated in separate studies for
internal consistency and test–retest reliability, accu-
racy, face validity, and clinical utility.
Internal consistency and test–retest reliability.
Two population-based surveys were conducted in de-
mographically diverse areas of Baltimore, Maryland
and Greater London, UK.
25,26
A representative sam-
ple of households was selected from both populations
using random-digit telephone dialing methods. Re-
spondents were interviewed about their headaches
using a clinically validated computer-assisted tele-
phone interview (CATI).
29
Individuals who received a
CATI diagnosis of migraine headache were invited to
participate in the reliability study. A total of 97 mi-
graine cases in the United States and 100 in the UK
completed the MIDAS Questionnaire on two occa-
sions, separated by a median interval of 21 or 22
days.
The goal of these two studies was to evaluate the
test–retest reliability of responses to specific MIDAS
questions and, more importantly, that of the overall
MIDAS score. A correlation of 0.8 or greater is in the
excellent to outstanding range.
25,26
Internal consistency of the MIDAS score was as-
sessed using the Cronbach a, a measure that is anal-
ogous to a split half reliability assessment. The
Cronbach a ranges from 0 to 1. A score of 0.7 is
considered acceptable, and 0.8 or more indicates that
the internal consistency of the scale is excellent.
25,26
However, Cronbach’s a is influenced by the number
of items that comprise a measure. Fewer items may
be limiting.
The number of days for which migraine sufferers
reported disability differed by activity domain. In
both the United States and the UK, more missed
days were reported for household work, followed in
order by non-work activities and employment (table
2).
25,26
Days during which productivity was reduced
by half or more in employment were considerably
more common than missed days of employment. In
contrast, the number of missed household work days
was similar to the number of days during which pro-
ductivity was reduced by half or more. The overall
mean MIDAS score was significantly higher in the
United States than in the UK, attributable, in part,
to a higher frequency of reported headaches in the
United States sample.
Overall, migraine sufferers were reliable in re-
porting information on lost time associated with
headaches (table 3). The test–retest Pearson correla-
tion coefficient responses for questions ranged from
0.54 to 0.68 in the United States and from 0.52 to
0.82 in the UK. Values for Spearman correlation co-
efficients were similar to the Pearson values. The
test–retest reliability of the overall MIDAS score was
approximately 0.8 in the United States and the UK.
Figure 1. The Migraine Disability Assessment (MIDAS)
Questionnaire.
S22 NEUROLOGY 56(Suppl 1) March 2001

This is an indication that MIDAS has excellent reli-
ability, especially given that there was an average
3-week interval between completion of the two
MIDAS Questionnaires. Real change in headache ex-
perience is likely to have occurred in some individu-
als and thus to have diminished the overall observed
reliability (table 3). The Cronbach a measure of in-
ternal consistency was 0.76 in the United States and
0.73 in the UK.
In the United States test–retest reliability study,
MIDAS scores from the first questionnaire adminis-
tered to the migraine cases were compared to those
of 80 control subjects without migraine but who re-
ported at least one headache in the previous year.
25
Migraine patients reported significantly more dis-
ability in all three domains than those with other
types of headache, especially in missed days of
household work and days of reduced productivity at
work. The mean MIDAS score for migraine cases
was more than threefold higher than that of controls
(table 4).
The overall MIDAS score is highly reliable. A
broad spectrum of migraine headache sufferers was
represented because the studies were population-
based. Parallel studies conducted in the United States
and the UK yielded highly consistent results. The dis-
tribution of MIDAS scores was shown to be substan-
tially and significantly different in migraine patients
compared with non-migraine headache sufferers.
Validity (accuracy). Accuracy of the MIDAS
score was evaluated by comparison with a 90-day
headache diary.
27
A population-based sample of mi-
graine sufferers in Baltimore, Maryland was re-
cruited and clinically assessed before being enrolled
in a 90-day study.
27
Study participants were exam-
ined by a physician to confirm migraine diagnosis
and were trained in the use of the daily diary. Partic-
ipants were instructed to complete a daily diary at
the same time each day, preferably at night. Each
diary booklet covered 1 week. On all days, whether
or not a headache occurred, questions were answered
about work, household work, other activities, and
stress and mood. On days with headache, informa-
tion was recorded on headache duration, pain inten-
Table 2 MIDAS scores in population studies from the USA and UK
25,26
Question
US UK
Mean Median Mean Median
1. On how many days in the past 3 months did you miss work or school
because of your headaches?
1.6 0 0.8 0
2. How many days in the past 3 months was your productivity at work or
school reduced by half or more because of your headaches?
4.9 2 3.9 2
3. On how many days in the past 3 months did you not do household work
because of your headaches?
5.3 3 3.3 3
4. How many days in the past 3 months was your productivity in household
work reduced by half or more because of your headaches?
4.7 3 3.8 2
5. On how many days in the past 3 months did you miss family, social or non-
work activities because of your headaches?
3.0 2 1.8 1
Total MIDAS score 19.5 14 12.8 9
Table 3 Results from a USA and UK study of test–retest reliability; Pearson and Spearman correlation coefficients for responses to
MIDAS questions and the overall score among participants in the USA (n 5 97) and UK (n 5 100)
25,26
Question
USA UK
Pearson
correlation
Spearman
correlation
Pearson
correlation
Spearman
correlation
1. On how many days in the past 3 months did you miss work or school because
of your headaches?
0.68* 0.69 0.66 0.56
2. How many days in the past 3 months was your productivity at work or
school reduced by half or more because of your headaches?
0.54 0.65 0.71 0.65
3. On how many days in the past 3 months did you not do household work
because of your headaches?
0.60 0.63 0.82 0.58
4. How many days in the past 3 months was your productivity in household
work reduced by half or more because of your headaches?
0.64 0.59 0.60 0.56
5. On how many days in the past 3 months did you miss family, social or non-
work activities because of your headaches?
0.62 0.71 0.52 0.46
Total MIDAS score 0.80 0.78 0.83 0.77
* p Values for all correlation coefficients are ,0.001.
March 2001 NEUROLOGY 56(Suppl 1) S23

sity, other pain features, associated symptoms,
productivity levels in work, household work, and
non-work activities. The “gold standard” or reference
measures for assessing validity of the MIDAS ques-
tions and the MIDAS score were equivalent mea-
sures derived from the 90-day diary data. Mean and
median values for MIDAS items and the MIDAS
score were compared to the diary-based reference
measures. Pearson and Spearman correlation coeffi-
cients were used to assess the extent to which the
MIDAS items and the MIDAS score were correlated
with equivalent diary measures.
27
A total of 144 subjects completed both the diary
study and a MIDAS Questionnaire at the end of the
diary period. The MIDAS items for missed days of
employment (Question 1), household work (Question
3), and non-work activities (Question 5) were similar
to diary-based estimates (table 5). MIDAS values for
reduced productivity in employment (Question 2)
and household work (Question 4) over-estimated
equivalent values derived from the diary. On aver-
age, the number of days with headache (Question A)
and average pain intensity (Question B) recorded on
MIDAS was similar to the diary-based measures.
27
Pearson correlation coefficients between individ-
ual MIDAS items and composite diary measures
ranged from 0.50 to 0.77 (table 6). The highest coef-
ficients were reported for average pain intensity,
number of days with headache, missed leisure days,
and missed days in employment. The correlation be-
tween the MIDAS score and the equivalent diary
measure was 0.63.
27
The results of the diary study indicate that
MIDAS provides a reasonably accurate composite
measure of headache-related disability. The validity
of the MIDAS score was superior to that of the
HImQ score, which was evaluated in a study using
the same design.
24
The correlations between MIDAS
Table 4 Results from the USA test–retest reliability study; mean MIDAS scores for migraine cases (n 5 97) and control subjects with
non-migraine headache (n 5 80). p Values test for t-test difference between the two groups
25
Question
Migraine cases
(n 5 97)
Controls
(n 5 80) p Value
1. On how many days in the past 3 months did you miss work or school because
of your headaches?
1.63 0.31 ,0.001
2. How many days in the past 3 months was your productivity at work or school
reduced by half or more because of your headaches?
4.93 1.49 ,0.001
3. On how many days in the past 3 months did you not do household work
because of your headaches?
5.27 1.60 ,0.001
4. How many days in the past 3 months was your productivity in household
work reduced by half or more because of your headaches?
4.67 1.90 ,0.001
5. On how many days in the past 3 months did you miss family, social or non-
work activities because of your headaches?
2.97 0.90 ,0.001
Total MIDAS score 19.46 6.20 ,0.001
Table 5 Results from the MIDAS validity study; mean and median values for MIDAS items and the MIDAS score compared with
equivalent composite measures derived from diary data
27
Question
MIDAS score
(mean)
Diary score
(mean)
MIDAS score
(median)
Diary score
(median)
1. On how many days in the past 3 months did you miss work
or school because of your headaches?
0.96 1.23 0.00 0.00
2. How many days in the past 3 months was your productivity
at work or school reduced by half or more because of your
headaches?
3.77 2.22 2.00 0.98
3. On how many days in the past 3 months did you not do
household work because of your headaches?
3.64 3.93 2.00 2.01
4. How many days in the past 3 months was your productivity
in household work reduced by half or more because of your
headaches?
3.92 2.94 2.00 1.06
5. On how many days in the past 3 months did you miss family,
social or non-work activities because of your headaches?
2.60 2.22 1.00 0.95
A. On how many days in the past 3 months did you have a
headache?
13.28 14.07 10.00 11.67
B. On a scale of 0–10, how painful were these headaches on
average?
6.06 5.71 6.00 5.65
Total MIDAS score 14.53 13.45 9.00 8.36
S24 NEUROLOGY 56(Suppl 1) March 2001

and diary-based measures were relatively strong,
given that measures were compared on the basis of
very different methods of collecting data. Although
responses to specific MIDAS questions about missed
days of activity were somewhat biased, the overall
MIDAS score itself was unbiased.
Face validity and clinical utility. The face valid-
ity and clinical utility of the MIDAS score were ex-
amined by determining whether MIDAS scores were
associated with physicians’ clinical judgments of
pain, disability, and need for medical care.
30
Twelve
cases were systematically sampled from among the
participants in a previous diary study.
24
Data from
the 90-day diaries were used to construct relevant
case histories. A group of 49 physicians was pre-
sented with the 12 case histories, one at a time, and
asked to rate the pain severity (mild, moderate, or
severe), degree of disability (none, mild, moderate or
severe) and the need for medical care (rated from 0
as lowest need to 100 as highest need) for each indi-
vidual. Responses were recorded on an electronic
keypad. Physicians were not aware of the MIDAS
scores for any case (which were based on the histo-
ries and ranged from 0 to 24), the components of the
score, or the purpose of the exercise.
After all 12 cases were presented, the MIDAS
Questionnaire and its scoring procedure were de-
scribed. Physicians were asked to evaluate the ease
of use of the MIDAS Questionnaire (rated as “very
easy,” “easy,” “not difficult,” “difficult,” or “very diffi-
cult”). They were also asked for the MIDAS scores
they would require to rate a patient’s activity limita-
tions or disability as mild, moderate, or severe, and
at which it was “not urgent,” “moderately urgent,”
and “very urgent” to prescribe the right treatment
for the patient at the time of initial consultation.
Summary measures (median, mean, and inter-
quartile ranges) were derived for the physicians’ rat-
ings of pain, disability, and medical need for each
case. Physician-graded results for pain and disability
levels and ease of use were recorded as percentages.
Within-physician correlation coefficients were de-
rived for the relationships between the MIDAS
scores and the physician-rated medical need for each
patient.
30
Forty-nine primary care physicians, neurologists,
and headache specialists from 14 countries took part
in the study. In general, the physician-rated median
intensity of pain and disability increased as the
MIDAS score increased (figure 2). The relationships
between the physician assessments of medical need
and the MIDAS score for each of the 12 case histories
are shown in figure 3. Each line represents a
regression-based estimate of this relationship for
each of the 49 physicians. In general, the physicians’
judgment of need for medical care increased as the
MIDAS score increased. The plot also demonstrates
a very high degree of coherence in perceived need for
medical care across physicians and MIDAS scores
(mean correlation coefficient 5 0.69). Eighty-nine
percent of the physicians stated that they believed
the MIDAS Questionnaire was “very easy,” “easy,” or
“not difficult” to complete.
30
The physicians’ judgments of the MIDAS score,
corresponding to different levels of a patient’s limita-
tion of activities (mild, moderate, or severe) and
treatment need (not urgent, moderately urgent, and
very urgent), were used to define the MIDAS score
into four grades of severity (table 7). The MIDAS
score for mild limitations to activity and low treat-
ment need was used to define the division between
MIDAS Grades I and II. The score for infrequent or
moderate limitations and moderate treatment need
was used to define the division between Grades II
and III. The score for severe limitations and severe
treatment need was used to define the division be-
tween Grades III and IV. These divisions were de-
fined so that they would be relatively easy to
remember and to ensure that non-disabled sufferers
Table 6 Results from the validity study; Pearson and Spearman correlation coefficients between MIDAS scores and equivalent measures
derived from the diary study
27
Question
Pearson
correlation
Spearman
correlation
1. On how many days in the past 3 months did you miss work or school because of your
headaches?
0.59 0.53
2. How many days in the past 3 months was your productivity at work or school reduced by
half or more because of your headaches?
0.55 0.42
3. On how many days in the past 3 months did you not do household work because of your
headaches?
0.50 0.53
4. How many days in the past 3 months was your productivity in household work reduced by
half or more because of your headaches?
0.54 0.59
5. On how many days in the past 3 months did you miss family, social or non-work activities
because of your headaches?
0.64 0.62
A. On how many days in the past 3 months did you have a headache? 0.70 0.71
B. On a scale of 0–10, how painful were these headaches on average? 0.77 0.76
Total MIDAS score 0.63 0.63
March 2001 NEUROLOGY 56(Suppl 1) S25

were not included in the groups with meaningful
disability.
30
MIDAS scores correlate with physician judgments
about pain, disability, and the need for medical care.
MIDAS appears to capture relevant information that
physicians use to make a clinical judgment about
migraine severity and treatment need. These results
support the face validity of the MIDAS Question-
naire. MIDAS was judged easy to use by most physi-
cians. The four MIDAS grades provide a
straightforward means of categorizing migraine pa-
tients according to their illness severity.
Discussion. The MIDAS Questionnaire was de-
signed to measure disability (activity limitations)
due to headache in three important domains of activ-
ity (employment, household work, and non-work ac-
tivities). MIDAS captures both actual missed days of
activity and days on which productivity is substan-
tially reduced. This approach focuses on important
domains of activity and provides a yardstick with
readily interpretable units (lost days) of disability.
The MIDAS Questionnaire was developed in re-
sponse to the practical limitations of the HImQ
16,24
and to meet a number of other needs, including facil-
itating communication in clinical practice, providing
a self-assessment procedure for screening individu-
als with severe disease who could benefit from medi-
cal care, and providing an outcome measure for
clinical practice, clinical trials, and epidemiologic
studies.
The studies described here show that the MIDAS
Questionnaire is brief and easy to complete and
score
30
and exhibits high test–retest reliability,
25,26
Figure 2. Relation between physician-rated median pain
intensity (a) and disability (b) with MIDAS score.
30
Figure 3. Relation between physician perception of medi-
cal need based on history and MIDAS score. Linear re-
gression lines for each of the 49 physicians based on their
evaluations of 12 patients.
30
Table 7 Grading system for the MIDAS Questionnaire
MIDAS Grade Definition Score
I Minimal or infrequent disability 0–5
II Mild or infrequent disability 6–10
III Moderate disability 11–20
IV Severe disability 211
S26 NEUROLOGY 56(Suppl 1) March 2001

reasonably good validity (accuracy),
27
and corre-
sponds to physicians’ judgments of illness severity
and medical need.
30
In addition, the MIDAS scores
differ substantially between migraine sufferers and
those with non-migraine headaches.
25
These many
strengths support the utility of MIDAS as an instru-
ment to assess headache severity both in the re-
search and the clinical practice setting.
3
Although the MIDAS score is a continuous mea-
sure, categorical groupings are likely to improve the
clinical utility of the measure.
31
There is a tradition
in medicine of defining categories to facilitate treat-
ment decisions, even for measures that have an un-
derlying continuous distribution. For example,
hypertension is defined in clinical practice based on
a single cut-score of 90 mm Hg in diastolic blood
pressure. The MIDAS grades define patients in four
disability categories: little or none, mild, moderate,
and severe. Studies show that roughly equal pro-
portions of migraine sufferers fall into each of the
MIDAS grades.
25,26
Because the MIDAS score separately assesses lost
time in each of three domains, a single day of dis-
abling headache may contribute more than one point
to the total score, depending on the duration and
severity of the attack and the range of activities
scheduled for that day. The MIDAS score therefore
reflects the range of intended activities lost on a
particular headache day summed over a 3-month pe-
riod. The studies show that migraine sufferers are
more likely to miss household work than non-work
activities and are more likely to miss non-work activ-
ities than their employment.
25,26
A very severe head-
ache is therefore likely to result in missed activity in
more than one domain. Because a single headache
day can contribute up to three points to the MIDAS
score, physicians must interpret this in light of the
total number of days with headache (provided in
Question A on the MIDAS Questionnaire).
Previous attempts to measure migraine severity
have used composite measures of pain and disabili-
ty.
16,23,24
MIDAS focuses on disability alone because
the aim was to produce a single uniform scale that
was easy to score and because information on dis-
ability, rather than level of pain, is likely to be the
dominant factor in influencing treatment decisions.
The MIDAS score is similar in some ways to
disease-specific HRQoL measures.
5-9
Like HRQoL
measures, MIDAS is intended to capture the aggre-
gate impact of illness on an individual over a period
of time. However, MIDAS differs from HRQoL in-
struments in design and content. HRQoL instru-
ments measure aspects of a migraine sufferer’s
health status over time, including days with and
without headache. MIDAS focuses only on days with
headache and is intended to measure severity of ill-
ness, expressed as a composite of lost time in various
domains, not HRQoL. HRQoL instruments tend to be
much longer than MIDAS and to have arbitrary or
percentile scores. MIDAS is currently the only in-
strument that has been shown to correlate with both
physicians’ assessment of treatment need and out-
comes of treatment.
In conclusion, the MIDAS Questionnaire is easy to
complete and score. It reliably and validly assesses
disability in the general population of migraine suf-
ferers and corresponds to physicians’ judgments of
illness severity and treatment needs based on com-
plete case histories. General use of the MIDAS Ques-
tionnaire has the potential to improve care for
migraine in clinical practice. It is particularly suited
to primary care because of its brevity and simplicity
of use and scoring.
Acknowledgment
The authors thank Dr. Peter Blakeborough for his help in the
preparation of this manuscript.
References
1. Lipton RB, Amatniek JC, Ferrari MD, et al. Migraine: identi-
fying and removing barriers to care. Neurology 1994;44(suppl
4):63–68.
2. Stewart WF, Shechter A, Lipton RB. Migraine heterogeneity.
Disability, pain intensity, and attack frequency and duration.
Neurology 1994;44(suppl 4):24–39.
3. Lipton RB, Goadsby PJ, Sawyer JPC, et al. Migraine: diagno-
sis and assessment of disability. Rev Contemp Pharmacother
2000;11:63–73.
4. Dahlo¨f C. Assessment of health-related quality of life in mi-
graine. Cephalalgia 1993;13:233–237.
5. Jacobson GP, Ramadan NM, Aggarwal SK, et al. The Henry
Ford Hospital Headache Disability Inventory. Neurology
1994;44:837–842.
6. Babiak LM, Miller DW, MacMillan JH, et al. Migraine-specific
quality of life: a comparison of US and Canadian results. Qual
Life Res 1994;3:58. Abstract.
7. Richard A, Henry P, Chezot G, et al. Quality of life and mi-
graine, validation of the QVM questionnaire in hospital con-
sultation and in general medicine [in French]. Therapie 1993;
48:89–96.
8. Solomon GD, Skobieranda FG, Graff LA. Quality of life and
well-being of headache patients: measurement by the Medical
Outcomes Study Instrument. Headache 1993;33:351–358.
9. Osterhaus JT, Townsend RJ, Gandek B, et al. Measuring the
functional status and well-being of patients with migraine
headache. Headache 1994;34:337–343.
10. Wagner TH, Patrick DL, Galer BS, et al. A new instrument to
assess the long-term quality of life effects from migraine: de-
velopment and psychometric testing of the MSQOL. Headache
1996;36:484–492.
11. National Academy of Sciences/Institute of Medicine (NAS/
IOM). Disability in America: toward a national agenda for
prevention. Washington, DC: NAS Press, 1991.
12. Janca A, Prilipko, Costa E, Silva JA. The World Health Orga-
nization’s global initiative on neurology and public health.
J Neurol Sci 1997;145:1–2.
13. Murray CJL, Lopez AD. The global burden of disease. Geneva:
World Health Organization, 1996.
14. Murray CJL, Lopez AD. Mortality by cause for eight regions
of the world: Global Burden of Disease Study. Lancet 1997;
349:1269–1276.
15. Murray CJL, Lopez AD. Regional patterns of disability-free
life expectancy and disability-adjusted life expectancy: Global
Burden of Disease Study. Lancet 1997;349:1347–1352.
16. Stewart WF, Lipton RB, Simon D, et al. Reliability of an
illness severity measure for headache in a population sample
of migraine sufferers. Cephalalgia 1998;18:44–51.
17. Schwartz BS, Stewart WF, Lipton RB. Lost workdays and
decreased work effectiveness associated with headache in the
workplace. J Occup Environ Med 1997;39:320–327.
18. Von Korff MR, Stewart WF, Simon DJ, et al. Migraine and
reduced work performance—a population-based diary study.
Neurology 1998;50:1741–1745.
19. Rasmussen B, Jensen R, Olesen J. Impact of headache on
March 2001 NEUROLOGY 56(Suppl 1) S27

sickness absence and utilisation of medical services: a Danish
population study. J Epidemiol Community Health 1992;46:
443–446.
20. de Lissovoy G, Lazarus SS. The economic cost of migraine.
Present state of knowledge. Neurology 1994;44(suppl 4):56–
62.
21. Slater ND, Lipton RB, Stewart WF, et al. Doctor–patient com-
munication about migraine disability. Neurology 1999;
52(suppl 2):A209. Abstract.
22. Lipton RB, Stewart WF, MacGregor A, et al. Communication
of migraine disability between physicians and patients. Ceph-
alalgia 1999;19:337. Abstract.
23. Von Korff M, Ormel J, Keefe FJ, et al. Grading the severity of
chronic pain. Pain 1992;50:133–149.
24. Stewart WF, Lipton RB, Simon D, et al. Validity of an illness
severity measure for headache in a population sample of mi-
graine sufferers. Pain 1999;79:291–301.
25. Stewart WF, Lipton RB, Kolodner K, et al. Reliability of the
migraine disability assessment score in a population-based
sample of headache sufferers. Cephalalgia 1999;19:107–114.
26. Stewart WF, Lipton RB, Sawyer J. An international study to
assess the reliability of the Migraine Disability Assessment
(MIDAS) score. Neurology 1999;52:988–994.
27. Stewart WF, Lipton RB, Kolodner KB, et al. Validity of the
Migraine Disability Assessment (MIDAS) score in comparison
to a diary-based measure in a population sample of migraine
sufferers. Pain 2000;88:41–52.
28. Scher AI, Stewart WF, Lipton RB. Migraine and headache: a
meta-analytic approach. In: Crombie IK, ed. Epidemiology of
pain. Seattle: IASP Press, 1999:159–170.
29. Lipton RB, Stewart WF, Liberman J, et al. Validation of a
computer assisted telephone interview for migraine (CATI-M):
results of independent validation studies in the US and the
UK. Headache 2000;40:417. Abstract.
30. Lipton RB, Stewart WF, Sawyer J, et al. Clinical utility of a
new instrument assessing migraine disability: the Migraine
Disability Assessment (MIDAS) Questionnaire. Headache.
Submitted.
31. Pilgrim AJ. Methodology of clinical trials of sumatriptan in
migraine and cluster headache. Eur Neurol 1991;31:295–299.
S28 NEUROLOGY 56(Suppl 1) March 2001