/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-97232-en.cckm

201606168

page

100

UWHC,UWMF,

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

Pittsburgh Sleep Quality Index - A New Instrument for Psychiatric Practice and Research

Pittsburgh Sleep Quality Index - A New Instrument for Psychiatric Practice and Research - Clinical Hub, UW Health Clinical Tool Search, UW Health Clinical Tool Search, Questionnaires, Related


Psychiatry Research. 28, 193-2 I3
Elsevier
193
The Pittsburgh Sleep Quality Index: A New Instrument for
Psychiatric Practice and Research
Daniel J. Buysse. Charles F. Reynolds Ill, Timothy H. Monk,
Susan R. Berman, and David J. Kupfer
Received May 9, 1988; revised version received August 17, 1988; accepted November 12, 1988.
Abstract. Despite the prevalence of sleep complaints among psychiatric patients,
few questionnaires have been specifically designed to measure sleep quality in
clinical populations. The Pittsburgh Sleep Quality Index (PSQI) is a self-rated
questionnaire which assesses sleep quality and disturbances over a l-month time
interval. Nineteen individual items generate seven “component” scores: subjective
sleep quality, sleep latency, sleep duration, habitual sleep efficiency, sleep
disturbances, use of sleeping medication, and daytime dysfunction. The sum of
scores for these seven components yields one global score. Clinical and clinimetric
properties of the PSQI were assessed over an l8-month period with “good”
sleepers (healthy subjects, n = 52) and “poor” sleepers (depressed patients, n = 54;
sleep-disorder patients, n = 62). Acceptable measures of internal homogeneity,
consistency (test-retest reliability), and validity were obtained. A global PSQI
score > 5 yielded a diagnostic sensitivity of 89.6% and specificity of 86.5% (kappa
= 0.75, p < O.OOl) in distinguishing good and poor sleepers. The clinimetric and
clinical properties of the PSQI suggest its utility both in psychiatric clinical
practice and research activities.
Key Words. Sleep, sleep quality, depression, sleep disorders.
“Sleep quality” is an important clinical construct for two major reasons. First,
complaints about sleep quality are common; epidemiological surveys indicate that
1535% of the adult population complain of frequent sleep quality disturbance, such
as difficulty falling asleep or difficulty maintaining sleep (Karacan et al., 1976, 1983;
Bixler et al., 1979; Lugaresi et al., 1983; Welstein et al., 1983; Mellinger et al., 1985).
Second, poor sleep quality can be an important symptom of many sleep and medical
disorders. One frequently measured component of sleep quality, sleep duration, may
even have a direct association with mortality (Kripke et al., 1979).
Sleep quality complaints are particularly relevant to psychiatry. Factors relating
to anxiety and stress are one of the most important concomitants of sleep complaints
in the general population (Karacan et al., 1983), and insomnia associated with
psychiatric disorders is the most prevalent type of insomnia seen in sleep disorders
centers, accounting for 35% of diagnoses (Coleman, 1983). Furthermore, sleep
Daniel J. Buysse, M.D., is Assistant Professor; Charles F. Reynolds III, M.D., is Professor; Timothy H.
Monk, Ph.D., is Associate Professor; Susan R. Berman is Research Associate; and David J. Kupfer,
M.D., is Professor and Chairman, Department of Psychiatry, University of Pittsburgh School of
Medicine. (Reprint requests to Dr. C.F. Reynolds Ill, Western Psychiatric institute and Clinic, University
of Pittsburgh, 381 I O’Hara St., Pittsburgh, PA 15213, USA.)
Ol65-1781/89/$03.50 @ 1989 El sevier Scientific Publishers Ireland Ltd.

194
quality disturbances are frequently reported in essentially all psychiatric disorders,
including depression, schizophrenia, anxiety disorders, and psychoactive substance
use disorders.
Although sleep quality is a readily accepted clinical construct, it represents a
complex phenomenon that is difficult to define and measure objectively. “Sleep
quality” includes quantitative aspects of sleep, such as sleep duration, sleep latency,
or number of arousals, as well as more purely subjective aspects, such as “depth” or
“restfulness” of sleep. However, the exact elements that compose sleep quality, and
their relative importance, may vary between individuals. Furthermore, because steep
quality is largely subjective, sleep laboratory measures may correlate with perceived
sleep quality, but they cannot define it. Finally, the measurement of sleep quality is
affected by the type of study in which it is being examined. Large-scale population
surveys generally focus on a few general questions about habitual sleep quality and
types of sleep disturbances (e.g., Bixler et al., 1979; Karacan et al., 1983). Studies
that examine the previous night’s sleep (drug efficacy studies, for example) may
focus on more subjective, comparative aspects of sleep quality, such as depth of
sleep, restfulness, and feelings upon awakening (e.g., Frankel et al., 1976; Webb et
al., 1976; Parrott and Hindmarch, 1978).
Given the importance of the construct and the inherent difficulties in its definition
and quantification, it is important to have a clinical instrument that measures sleep
quality. It is also necessary, however, to assess the “clinimetric” properties (i.e.,
properties such as sensibility, accuracy, comprehensibility, and reproducibility) of
the instrument, all of which are essential to the description and valid measurement of
complex clinical phenomena (Feinstein, 1987). Although many sleep questionnaires
have been described in previous studies, they share several general difficulties. First,
very few of them have used specified time intervals for assessment. Second, previous
questionnaires have not been designed to yield a simple, global score to facilitate
comparisons between groups or individuals. Third, few of these studies have directly
assessed clinimetric properties of the questionnaires. Finally, previous question-
naires have been used primarily with unselected population samples or nonclinical
control subjects.
The Pittsburgh Sleep Quality Index was developed with several goals: (I) to
provide a reliable, valid, and standardized measure of sleep quality; (2) to
discriminate between “good” and “poor” sleepers; (3) to provide an index that is easy
for subjects to use and for clinicians and researchers to interpret; and (4) to provide a
brief, clinically useful assessment of a variety of sleep disturbances that might affect
sleep quality. This article describes the instrument and its clinimetric properties,
including internal homogeneity, performance consistency, and validity.
Methods
Instrument Development and Description (Appendix). Items in the Pittsburgh Sleep
Quality Index (PSQI) were derived from three sources: clinical intuition and experience with
sleep disorder patients; a review of previous sleep quality questionnaires reported in the
literature; and clinical experience with the instrument during 18 months of field testing.
The PSQI assesses sleep quality during the previous month. This is a time frame inter-
mediate between postsleep inventories (which assess only the previous night’s sleep) and

195
survey-type questionnaires (which assess difficulties over the previous year or more). A
postsleep questionnaire may reflect more accurately the night-to-night variations that occur in
sleep quality, but it does not provide information about the frequency or duration of specific
problems that may lead a patient to seek help. On the other hand, survey-type questionnaires
may not indicate the severity of a particular problem at the present time. In addition, a
duration of 2-3 weeks is often used clinically to differentiate transient from persistent sleep-
wake disorders (Consensus Conference on Insomnia, 1984). Therefore, administering the
PSQI on two occasions separated by approximately I month allows for the discrimination of
most transient and persistent disturbances.
The PSQI consists of 19 self-rated questions and five questions rated by the bedpartner or
roommate. The latter five questions are used for clinical information only, are not tabulated in
the scoring of the PSQI, and are not reported on in this article. The 19 self-rated questions
assess a wide variety of factors relating to sleep quality, including estimates of sleep duration
and latency and of the frequency and severity of specific sleep-related problems. These I9
items are grouped into seven component scores, each weighted equally on a O-3 scale. The
seven component scores are then summed to yield a global PSQI score, which has a range of
O-2 I; higher scores indicate worse sleep quality.
The seven components of the PSQI are standardized versions of areas routinely assessed in
clinical interviews of patients with sleep/ wake complaints. These components are subjective
sleep quality, sleep latency, sleep duration, habitual sleep efficiency, sleep disturbances, use of
sleeping medications, and daytime dysfunction. Scoring of each component is illustrated in
the Appendix. Subject instructions for the PSQI are contained in the text. The entire index
requires 5-10 min for the subject to complete, and 5 min to score.
Subjects. The PSQl was administered to three groups of subjects during an l&month study
period. Group I consisted of “good” sleepers: 52 healthy control subjects without sleep
complaints, recruited for participation in research studies of sleep and aging (MH-37869),
nocturnal penile tumescence (MH-40023), and sleep in depression (MH-40023, MH-30915).
Group 2 consisted of “poor” sleepers: 34 patients with major depressive disorder, who were
again recruited for participation in research protocols relating to sleep, aging, depression, and
nocturnal penile tumescence. This group included 24 outpatients and IO inpatients at the
Western Psychiatric Institute and Clinic. Group 3, also consisting of “poor” sleepers, was a
clinical sample of 62 physician-referred outpatients at the Sleep Evaluation Center (SEC) of
the Western Psychiatric Institute and Clinic. Patients are referred to the SEC for assessment
of a variety of sleep/wake complaints, but only patients with Disorder of Initiating and
Maintaining Sleep (DIMS, n = 45) or Disorders of Excessive Somnolence (DOES, n = 17)
(Association of Sleep Disorders Centers-ASDC, 1979) were included in this study, since the
number of patients with other disorders was too small to permit statistical analysis.
Subjects were not matched for age or sex ratio because of the different requirements for
each research protocol, and the absence of any age criteria for the clinical sleep disorders
sample. The mean ages for subject groups were as follows: controls 59.9 years (range: 24-83);
depressives 50.9 years (range: 21-80); DIMS 44.8 years (range: 20-80); and DOES 42.2 years
(range: 19-57). Analysis of variance (ANOVA) indicated a significant difference in age
between groups (F= 5.20,~ < O.OOI), with post hoc differences between control subjects and
DIMS and DOES patients. Male/female ratios were as follows: controls 40/ 12; depressives
25/9; DIMS l6/29; and DOES 8/9 b2 = 21.2,~ < O.OOl). Male subjects had a lower mean
age (46.5 years; SD = 16.7) than female subjects (55.4 years; SD = 18.9) (t = -3.01 ,p < 0.005).
Many of the male subjects were involved in studies of nocturnal penile tumescence in
depression, while female subjects were participating mainly in studies of sleep, aging, and
depression.
Evaluation for all subjects included a complete medical history and physical examination.
Depressives and controls were excluded from research involvement (and therefore, from the
current study) for any medical conditions that would prevent a 2-week medication-free
interval, as well as for the presence of known central nervous system disease such as

196
seizure disorder, cerebrovascular disease, or dementia. No specific exclusion criteria were used
for the clinic sample of sleep-disorder patients. All depressed patients and healthy controls
were assessed with the Schedule for Affective Disorders and Schizophrenia-Lifetime version
(SADS-L) (Endicott and Spitzer, l978), and diagnosed according to Research Diagnostic
Criteria (Spitzer et al., 1978); all depressed patients met criteria for definite or probable
current major depressive disorder. Severity of depressive symptoms was assessed with the
Hamilton Rating Scale for Depression (Hamilton, 1960); the mean Hamilton score for
depressed patients was 21.3 (SD = 4.65). Sleep-disorder patients were evaluated as described
elsewhere (Jacobs et al., 1988) and given preliminary diagnoses according to ASDC nosology
(ASDC, 1979). Sleepdisorder patients meeting criteria for DSM-III (American Psychiatric
Association, 1980) major depression were excluded from the current study. All subjects
completed a 2-week sleep/ wake diary and a sleep habits questionnaire.
All subjects were further evaluated with routine polysomnography following a medication-
free interval of at least 2 weeks. For depressed and sleepdisorder patients, this interval
followed withdrawal from psychotropic and sedative-hypnotic medications. All subjects were
studied with a routine sleep montage, including electroencephalographic (C4, referenced to
tied mastoids), electro-oculographic (EOG), and electromyographic (submental) leads. Most
subjects had additional monitoring for sleep apnea, myoclonus, or nocturnal penile
tumescence, dictated by clinical indications or research protocol involvement. All sleep
records were scored in I-min epochs according to standard criteria (Rechtschaffen and Kales,
1968) using Stage 2 sleep onset, and standard convention for definition of sleep efficiency
(time spent asleep/ total recording period).
Final diagnoses for depressed and sleep-disorder patients were based on results of clinical
and structured interviews, sleep questionnaires, and diaries. In addition, polysomnographic
findings were considered in the final diagnoses of the sleepdisorder patients (Jacobs et al.,
1988).
All I48 subjects completed the PSQI on at least one occasion (T,) during the course of their
clinical and research evaluation. For the majority of subjects (n = 107) the PSQI was
completed before sleep studies. For some subjects with stable sleep/ wake complaints (n = 41)
the PSQI was completed after sleep studies. A subgroup of 91 subjects (43 controls, 22
depressives, and 26 sleep-disorder patients) completed the index a second time (T,), an
average of 28.2 days later (range: l-265 days). The second PSQI was completed before any
pharmacological treatment began.
Statistical Analyses. Descriptive statistics and ANOVA were used to contrast clinical and
demographic features of the patient groups.
Internol homogeneity of separate items was assessed using Cronbach’s (x statistic and
corrected component-total correlation coefficients (Cronbach, 1951). Pearson product-
moment correlations were also used to correlate component and item scores with the PSQI
global score.
Test-retest reliability (consistency) was assessed with paired t tests and Pearson product-
moment correlations for PSQI global score, component scores, and individual items, at Time
I (T,) versus Time 2 (T,). This was done for the entire subject pool, as well as for separate
subject groups (except DOES patients, since only ftve patients had complete questionnaires on
two occasions).
As the primary analysis of validity, we assessed the degree to which the index detected
differences between groups recognized clinically as distinct. This assumes that the index
measures differences between groups at the same time point as a clinical “gold standard.” In
this case, the relevant “gold standard” diagnoses were based on a combination of clinical
interviews, structured interviews, and polysomnographic data. For this analysis, an analysis of
covariance (ANCOVA) was used to compare patient groups for PSQI global and component
scores, and the Student-Neuman-Keul’s procedure was used for pairwise comparisons. Age
and sex were used as covariates because of group differences in age and sex ratio. A multiple
ANCOVA (MANCOVA) was performed for the PSQI global score, again using age and sex
as covariates.

197
As a secondary analysis of validity, we compared PSQI scores with polysomnographic
results, being cognizant of the fact that PSQI scores reflect the experience of sleep during the
previous month, while polysomnographic data were limited to 2 or 3 nights. PSQI estimates of
sleep latency, sleep duration, and sleep efficiency were compared to their homologous
polysomnographic measures, using both t tests and Pearson product-moment correlations.
Global PSQI scores were also compared to polysomnographic variables selected a priori as
being likely to correlate with overall sleep quality, again using Pearson correlations. The
specific variables selected were REM Yc, delta 70, sleep latency, sleep efficiency, and sleep
duration. Finally, group differences for these polysomnographic variables were assessed using
one-way ANOVAs.
Results
General Results. Subjects found the PSQI easy to use and understand. Ten
subjects out of an original pool of 158 failed to give complete responses to all items,
and were therefore omitted from any further analyses; nine of these 10 were DOES
patients.
The PSQI global score has a possible range of O-21 points. Actual scores ranged
from 0 to 20 points, with an overall group mean of 7.4, median of 6.0, and SD of 5. I.
For individual components, each with a possible range of O-3, the observed ranges
were O-3.
Age was negatively correlated with the subjective sleep quality (r = -0.22, p < 0.05)
and daytime dysfunction (r = -0.29, p < 0.02) component scores in the healthy
controls. The PSQI global score and other component scores (sleep latency, sleep
duration, habitual sleep efficiency, sleep disturbances, and use of sleeping
medications) were not significantly correlated with age.
Internal Homogeneity. The seven component scores of the PSQI had an overall
reliability coefficient (Cronbach’s a) of 0.83, indicating a high degree of internal
consistency. In other words, each of the seven components appears to measure a
particular aspect of the same overall construct, viz., sleep quality. The largest
component-total correlation coefficients were found for habitual sleep efficiency and
subjective sleep quality (0.76 for each), and the smallest correlation coefficient was
found for sleep disturbances (0.35). The mean component-total correlation
coefficient was 0.58. Pearson product-moment correlations between component
scores and the PSQI global score were also calculated for the entire group, as well as
each group separately (Table I). Once again, the strongest correlations were seen for
habitual sleep efficiency and subjective sleep quality.
Individual items were also strongly correlated with each other, indicated by a
reliability coefficient (Cronbach’s a) of 0.83. Item-total correlation coefficients
ranged from 0.66 for question #9 (enthusiasm to get things done) to 0.20 for item #8
(difficulty staying awake). Pearson product-moment correlations between individual
items and the global score ranged from 0.83 (subjective sleep quality) to 0.07 (cough
or snore during sleep) (Table 2).
Performance Consistency (Test-Retest Reliability). Ninety-one patients
completed the PSQI on two separate occasions. Paired t tests for the global PSQI
score, as well as the seven individual component scores, showed no significant

198
differences between T, and T,. Two differences were noted for depressed patients,
who showed a reduction in sleep disturbances (t = 2.32, p = 0.03) and daytime
dysfunction (I = 3.46, p = 0.002) at T,.
Table 1. Component-global Pittsburgh Sleep Quality Index (PSQI) score
correlations’
All groups Controls Depressives DIMS2 DOES3
(n = 48) (n = 52) (n = 34) (n = 45) (n = 17)
Component r p r p r p r p r p
Sleep quality 0.83 0.001 0.64 0.001 0.71 0.001 0.68 0.001 0.57 0.01
Sleep latency 0.72 0.001 0.58 0.001 0.64 0.001 0.67 0.001 0.69 0.001
Sleep duration 0.80 0.001 0.44 0.001 0.68 0.001 0.79 0.001 0.60 0.01
Habitual sleep efficiency 0.85 0.001 0.57 0.001 0.83 0.001 0.75 0.001 0.76 0.001
Sleep disturbance 0.46 0.001 0.70 0.001 0.19 - 0.31 0.02 0.38 -
Use of sleeping
medication 0.62 0.001 0.20 - 0.69 0.001 0.51 0.001 0.33 -
Daytime dysfunction 0.63 0.001 0.53 0.001 0.38 0.01 0.53 0.001 0.38 0.001
1. Pearson product-moment correlations.
2. DIMS = Disorders of Initiating and Mamtaining Sleep
3. DOES = Disorders of Excessive Somnolence.
Pearson product-moment correlations again demonstrated stability in global and
component scores. For the entire subgroup in which T, and T, measures were
obtained, the T,-T, correlation coefficient for global PSQI scores was 0.85
@ < 0.001). Component scores had coefficients ranging from 0.84 (sleep latency) to
0.65 (medication use) @ < 0.001 for each component score). Global PSQI scores for
each diagnostic group were also significantly correlated between the two testing
times, with r’s > 0.40 @ < 0.005) for each group. Component scores within each
subject group showed more variability across time, but all of these scores were
significantly correlated (r’s > 0.35, p < 0.05). The single exception was medication
use in control subjects, which showed no correlation between the two testing times.
Validity. (Table 3, Figs. I, 2). Global PSQI scores differed significantly between
subject groups, using an ANCOVA with age and sex as covariates (Table 3). Control
subjects differed from all patient groups (Student-Neuman-Keul’s procedure).
Furthermore, DIMS and depressed patients had significantly higher scores than
DOES patients. Control subjects differed from DIMS and depressed patients on all
individual component scores; controls also differed from DOES patients on three
components (sleep disturbances, daytime dysfunction, and sleep quality). DOES and
DIMS patients had significantly different scores on all components except sleep
disturbances, and DOES and depressives patients differed on all components except
sleep disturbance and daytime dysfunction.
Group differences resulted in distinctive component and global score profiles,
shown in Fig. I. Depressed and DIMS patients showed similar profiles, which
differed from those of DOES patients and control subjects. These differences were
further substantiated with a significant MANCOVA for component scores across
groups (Hotelling’s TL = 2.62, p < 0.001).

199
Age was a significant covariate only for the daytime dysfunction component; but
contrary to expectations, these factors were inversely correlated, i.e., reported
severity of daytime dysfunction tended to be greater in younger than in older
subjects. Sex was a significant covariate for use of sleeping medications and habitual
sleep efficiency, with males showing higher scores for each of these components. Age
and sex were both significant covariates for the PSQl global score, but group
differences were highly statistically significant even after covarying for these factors.
The distribution of global PSQI scores also differed between groups (Fig. 2). A
post hoc cutoff score of 5 correctly identified 88.5% (131/ 148) of all patients and
controls (kappa = 0.75, p < 0.001). This represents a sensitivity of 89.6% and a
specificity of 86.5%. The same cutoff score correctly identified 84.4% (38/45) of
DIMS patients, 88% (IS/ 17) of DOES patients, and 97% (33/34) of depressives.
Group differences in PSQI global scores were also substantiated by
polysomnographic results, which showed significant group differences for sleep
latency (F= 4.53, p < O.OOl), sleep efficiency (F = 5.78, p < O.OOl), sleep duration
(F= 4.82,~ < 0.003), and number of arousals (F= 2.87,~ < 0.04). Significant group
differences were not found for rapid eye movement (REM) % or delta sleep %.
Validity of the PSQI was further examined by comparing PSQI estimates of sleep
variables with those obtained by polysomnography. T tests showed no differences
between PSQI estimates and laboratory findings for sleep latency, but PSQl
estimates of the past month’s usual sleep duration and efficiency were greater than
Table 2. Item-global Pittsburgh Sleep Quality Index (PSQI) score
correlations1
Item2
Q5A
Q5B
Q5C
Q5D
Q5E
Q5F
Q5G
Q5H
051
Q5J
08
Q6
Q2
Q4
Q7
Q9
All groups Controls Depressives DIMS3 DOES’
(n = 148) (n = 52) (n = 34) (n = 45) (n = 17)
r p r p r p r p r p
0.71 0.001 0.63 0.001 0.56 0.001 0.64 0.001 0.68 0.001
0.52 0.001 0.52 0.001 0.38 0.01 0.49 0.001 0.55 0.01
0.24 0.001 0.29 0.02 0.31 0.04 0.33 0.01 0.14 -
0.17 0.02 0.12 - 0.07 - 0.08 - -0.07 -
0.07 - 0.34 0.007 -0.17 - -0.10 - 0.29 -
0.29 0.001 0.03 - 0.03 - 0.23 - -0.03 -
0.18 0.01 0.26 0.03 0.02 - 0.07 - -0.00 -
0.20 0.007 0.31 0.01 0.25 - -0.21 - -0.03 -
0.24 0.002 0.54 0.001 0.06 - -0.02 - 0.33 -
0.32 0.001 0.37 0.004 0.17 - 0.22 - -0.13 -
0.19 0.009 0.21 - -0.08 - 0.28 0.03 0.14 -
0.83 0.001 0.37 0.003 0.71 0.001 0.69 0.001 0.57 0.008
0.66 0.001 0.64 0.001 0.56 0.001 0.64 0.001 0.72 0.001
0.80 0.001 0.44 0.001 0.68 0.001 0.79 0.001 0.57 0.008
0.62 0.001 0.19 - 0.69 0.001 0.49 0.001 0.31 -
0.69 0.001 0.55 0.001 0.44 0.005 0.53 0.001 0.30 -
I. Pearson product-moment correlations.
2. Refer to questionnaire in Appendix.
3. DIMS = Disorders of Mating and Maintaining Sleep.
4. DOES = Disorders of Excessive Somnolence.

Ta
bl
e
3.
P
it
ts
bu
rg
h
Sl
ee
p
Qu
al
ity
In
de
x
(P
SQ
I)
co
mp
ar
iso
ns
b
et
we
en
di
ag
no
st
ic
g
ro
up
&
Me
an
sc
or
e
f

SD
(a
dj
us
te
d
m
ea
n)

Co
nt
ro
ls

De
pr
es
si
ve
s
DI
MS

DO
ES

AN
CO
VA

Si
gn
ifi
ca
nt
S
tu
de
nt
-N
eu
ma
n-

Co
mp
on
en
t
(
n

=

52
)
(
n

=

34
)
(
n

=

45
)
(n
=1
7)

F

D

Ke
ul
s’
c
om
oa
ris
on
s
Su
bj
ec
ti
ve
s
le
e
p
q
ua
lit
y
Sl
ee
p
la
te
nc
y
Sl
ee
p
d
ur
at
io
n
H
ab
itu
al
s
le
e
p
e
ff
ic
ie
n
cy

Sl
ee
p
d
is
tu
rb
an
ce
s
U
se
o
f s
le
e
p
in
g
m
ed
ica
tio
n
Da
yt
im
e
dy
sf
un
ct
io
n
PS
QI

g
lo
b
al
s
c
o
re

0.
35
f
0.
48

(
0
.4
0
)
0.
56
z
t
0.
73

(
0
.7
0
)
0.
29
f
0.
50

(
0
.3
1
)
0.
10
i
0.
30

(
0
.1
1
)
1 .
oo
f

0.
40

(
0
.9
5
)
0.
04
f
0.
28

(
0
.1
2
)
0.
35
f
0.
48

(
0
.4
4
)
2.
67
f
1.
70

1.
88
?
c
0
.8
8
(
1
.9
2
)
1.
88
f
1.
15

(
1
.9
6
)
1.
71
f
1.
14

(
1
.7
4
)
1.
59
+
1
.1
8
(
1
.6
3
)
1.
47
*
0.
51

(
1
.4
5
)
0.
76
+
1
.2
1
(
0
.8
4
)
1.
79
f
0.
69

(
1
.8
3
)
11
.O
Q
f
4.
31

1.
96
f
0.
93

(
1
.9
1
)
1.
42
f
1.
01

(
1
.3
1
)
1.
51
*
1.
20

(
1
.4
6
)
1.
47
f
1.
24

(
1
.4
1
)
1.
40
+
0
.6
2
(
1
.4
3
)
1.
20
+
1
.3
1
(
1
.O
Q
)
1.
42
k

0.
94

(
1
.3
7
)
IO
.3
8
f
4.
57

1.
06
f
0.
75

(
1
.0
2
)
0.
59
z
t
0.
87

(
0
.4
9
)
0.
47
f
0.
80

(
0
.4
6
)
0.
29
f
0.
77

(
0
.3
0
)
1.
53
f
0.
72

(
1
.5
6
)
0.
35
*

1.
00

(
0
.3
1
)
2.
24
f
0.
90

(
2
.1
6
)
6.
53
+
2
.9
8
35
.9

0.
00
01

15
.3

0.
00
01

20
.4

0.
00
01

25
.2
2
0.
00
01

8.
4
0.
00
01

7.
92

0.
00
01

33
.2
3
0.
00
01

45
.1
4
0.
00
1
C
on
tr
ol
s
vs
.
d
ep
re
ss
iv
es
, D
IM
S,
D
OE
S;

DO
ES
v
s.
d
ep
re
ss
iv
es
, D
IM
S
C
on
tr
ol
s
vs
. d
ep
re
ss
iv
es
,
DI
MS
;
DO
ES
v
s.
d
ep
re
ss
iv
es
,
DI
MS

C
on
tr
ol
s
vs
.
d
ep
re
ss
iv
es
,
DI
MS
;
DO
ES
v
s.
d
ep
re
ss
iv
es
,
DI
MS

C
on
tr
ol
s
vs
.
d
ep
re
ss
iv
es
, D
IM
S;

DO
ES
v
s.
d
ep
re
ss
iv
es
, D
IM
S
C
on
tr
ol
s
vs
. d
ep
re
ss
iv
es
,
DI
MS
, D
OE
S
C
on
tr
ol
s
vs
. d
ep
re
ss
iv
es
,
DI
MS
;
DO
ES
v
s.
d
ep
re
ss
iv
es
, D
IM
S
C
on
tr
ol
s
vs
. d
ep
re
ss
iv
es
,
DI
MS
, D
OE
S;

DO
ES
v
s.
D
IM
S
C
on
tr
ol
s
vs
. d
ep
re
ss
iv
es
,
DI
MS
, D
OE
S;

DO
ES
v
s.
d
ep
re
ss
iv
es
, D
IM
S
1.
DI
MS
=
D
is
or
de
rs
of
In
it
ia
ti
ng

an
d
Ma
in
ta
in
in
g Sl
ee
p.
DO
ES
=
D
is
or
de
rs
of
E
xc
es
si
ve
S
om
no
len
ce
.
2.
S
ig
n
if
ic
an
t ef
fe
ct
of
s
ex
a
s
co
va
ri
at
e.

3.
S
ig
n
if
ic
an
t ef
fe
ct
of
a
ge
a
s
co
va
ri
at
e.

4.
S
ig
n
if
ic
an
t ef
fe
ct
of
a
ge
a
nd
se
x
a
s
co
va
ri
at
es
.

201
Fig. 1. Pittsburgh Sleep Quality Index (PSQI): Mean component score
profiles
V
0
0
0
-I2
-10
-0
-0
-4
Depressed patients and patients with Disorders of Initiating and Maintaining Sleep (DIMS) have different components
score profiles than do control subjects. Patients with Disorders of Excessive Somnolence [DOES) have a profile more
similar to controls, but with expected elevations in subjective sleep quality. sleep disturbances, and daytme
dysfunction. Significant group differences for individual components and overall profiles were substantiated with
analyses of variance and muftiile analyses of covariance (Table 3).
those obtained during polysomnography (t = 9.98 and 4.50, respectively; both
p’s < 0.001). This pattern was true for the total subject pool as well as individual
subject groups. Pearson correlations demonstrated no significant positive
correlations between PSQI estimates and polysomnographic results, except in sleep
latency for the total subject pool (r = 0.33, p < 0.001) and for the depressive
subgroup (r = 0.37, p < 0.02). Similarly, the global PSQI score was compared with
several polysomnographic measures which we selected a priori as being likely to
correlate with perceived sleep quality. For all subjects, the global score was weakly
correlated only with objective sleep latency (r = 0.20, p < 0.01). For individual
subject groups, the global PSQI score correlated only with REM Yc in controls
(r = 0.34, p < 0.006) and number of arousals in depressives (r = 0.47, p < 0.002).
Discussion
Eighteen months of field testing with the PSQI have demonstrated that (1) subjects
and patients find the index easy to use; (2) the seven major components of the index,
as well as the 19 individual questions, are internally consistent; (3) the global scores,
component scores, and individual question responses are stable across time; (4) the
validity of the index is supported by its ability to discriminate patients from controls,
and, to a more limited degree, by concurrent polysomnographic findings. We will

202
Fig. 2. Pittsburgh Sleep Quality Index (PSQI) global scores
Healthy Controls (~52)
20
16
16
14
12
IO
6
6
4
2
0
0 1 2 3 4 5 6 7 6 91011~21314151617l6192021
Depressives (n=34)
10
6
6
1
2
0
0 1 2 3 4 5 6 7 6 9101112131416161716192021
Sleep Disorder Patients (n=62)
20-
16 -
16.
14 -
12 -
lo-
6-
6-
4-
2
L-
_,_.('
0
0
PSQI Global Score
PSQI global scores showed different distnbutions for control subjects, depressed patients, and sleep-disorder patients.
A global score cutoff of > 5 correctly identified 88.5% of all controls and patients, yielding a sensitivity of 89.6% and a
specificity of 86.5% (Kappa = 0.75. p < 0.001).
further discuss the format and clinimetric properties of the PSQI in relation to
previous sleep questionnaires in the literature. We will also discuss possible
applications for the PSQI in psychiatric clinical practice and research studies.
Questionnaire Format. A number of previous studies have reported on the use of
self-rated or interviewer-administered sleep-quality questionnaires. In general, these
instruments are of three types: habitual (i.e., “usual”) sleep questionnaires for
population surveys; habitual (“usual”) sleep questionnaires for clinical investi-
gations; and postsleep inventories.
The first type of questionnaire is used in epidemiological surveys of habitual or
usual sleep habits, sleep difficulties, and sleep quality (e.g., McGhie and Russell,

203
1962; Karacan et al., 1976, 1983; Bixler et al., 1979; Johnson and Spinweber, 1983;
Lugaresi et al., 1983; Welstein et al., 1983; Mellinger et al., 1985). The questions are
usually few in number and general in scope, typically focusing on sleep duration, the
presence of insomnia, and the use of medications for sleep. Habitual sleep
questionnaires have also been used in clinical studies, most often to compare
subjective reports with polysomnographic correlates (e.g., Monroe, 1969; Baekeland
and Roy, 1971; Mendelson et al., 1984, 1986) or to examine differences between
groups of subjects (McGhie, 1966; Beutler et al., 1978; Domino et al., 1984). These
questionnaires are often more detailed than those used in large-scale surveys, and
they include subjective estimates of sleep quality; however, their main focus is again
on quantitative measures. The final type of questionnaire found in the literature is
postsleep inventories (e.g., Samuel, 1964; Lewis, 1969; Frankel et al., 1973, 1976;
Carskadon et al., 1976; Webb et al., 1976; Parrott and Hindmarch, 1978, 1980; Ellis
et al., 198 I; Mendelson et al., 1984). These instruments ask a variety of quantitative
and qualitative questions about the previous night’s sleep. They vary considerably in
format, and in the number and type of questions. Postsleep inventories have been
used to examine differences between subjective reports and objective polysomno-
graphic findings, to study “good” and “bad” sleep, and to assess medication effects
on sleep.
The PSQI has some similarities to these other questionnaires, but also has some
important differences. The first comparison is in time interval of assessment. Most
habitual sleep questionnaires do not specify a particular time frame, although there
are some exceptions to this generalization (e.g., McGhie, 1966; Mendelson et al.,
1986). The PSQI assesses a l-month interval, which, as mentioned above, is clinically
and scientifically useful. While postsleep inventories are unambiguous in their
assessment of a single night’s sleep, they are not as useful for detecting patterns of
dysfunction, as noted previously.
A second comparison regards the type of questions included in the questionnaire.
The PSQI is similar to many of the habitual sleep questionnaires in the type of
questions included, e.g., estimates of sleep latency and duration, and frequency and
severity estimates of problems. The PSQI’s combination of quantitative and
qualitative information is not found, however, in some of the more carefully studied
questionnaires, such as those of Domino et al. (1984) and Webb et al. (1976).
The use of “component” scores in the PSQI is also similar to several other
questionnaires, which have generated between 4 and 11 “factors” relating to sleep
quality (Webb et al., 1976; Beutler et al., 1978; Parrott and Hindmarch, 1978;
Domino et al., 1984). One major difference is that other questionnaires have more
often included factors concerning mental activity before and during sleep. Another
difference is that these other questionnaires have used factor analysis to generate
specific factors, while the PSQI components are empirical and clinical in origin,
rather than statistical.
A third comparison between the PSQI and other questionnaires regards scoring
methods. The PSQI assigns ordinal scores to quantitative and qualitative
information, allowing for the generation of component scores and a single global
score. Except for McGhie (1966) and Beutler et al. (1978), previous questionnaires

204
do not use numerical scores for components or global scores. In the latter
questionnaire, standard scores were determined by transforming the actual values of
eight differently weighted “factors,” and assigning an arbitrary value of “50” to the
control mean. The PSQI global score has the advantages of giving a single overall
assessment of sleep quality, being simple to calculate, and allowing for direct
comparisons of individual patients or groups.
Finally, the PSQI was designed to assess clinical samples, while most previous
questionnaires have been designed to assess normal sleep habits or entire
populations. Although some questionnaires have been applied to patients with
insomnia diagnosed by ASDC criteria (Mendelson et al., 1984, 1986) most have
used patient samples that were not diagnosed according to current sleep disorders or
psychiatric nomenclature.
Clinical Properties: Homogeneity and Consistency. The Cronbach’s a of 0.83
obtained for PSQI components indicates a high degree of internal homogeneity
(Feinstein, 1987). In other words, each of the components measures part of a
coherent overall construct. Subjective sleep quality (item #6) showed one of the
highest component-total correlation coefficients, which further supports this notion.
The low component-total correlation coefficient seen for sleep disturbances may be
the result of the large number of items that make up this component, as well as the
fact that these items may be particularly susceptible to variation between individuals
and over time. The sleep disturbance component also showed the least difference
between diagnostic groups.
Data about internal homogeneity have been reported for three other sleep
questionnaires, each of which used factor-analytic techniques. Domino et al. (1984)
reported a factor analysis which yielded seven factors accounting for 7 1.7% of the
total variance in the questionnaire. Cronbach’s a was computed for these factors,
with a median a of 0.82. Beutler et al. (1978) analyzed their 186-item questionnaire
and found eight factors accounting for 59% of the total variance. The number of
questions contributing to each factor ranged between 4 and 17. Finally, the postsleep
inventory of Webb et al. (1976) had seven factors accounting for 54.5% of the total
variance; specific questions with inter-item correlations higher than 0.65 were
deliberately excluded. It is difficult to compare the PSQI with these questionnaires,
due to differences in time frame of assessment, number of total questions, and
subject populations tested. Most important, components of the PSQI were selected
on purely clinical grounds, and not on the basis of factor analysis.
Overall consistency (test-retest reliability) of the PSQI was better for the entire
subject pool than for any specific group. Of particular interest is the finding that
DIMS patients had the highest correlations across time, while the control subjects
had the lowest correlations. One possible explanation for the lower stability in the
control subjects’ scores is their low scores on all components, and particularly
medication use. For both control subjects and DIMS patients, component
correlations were highest for sleep latency and sleep duration, two “quantitative”
variables. Correlation coefficients were lower for daytime dysfunction and sleep
quality, two of the more “qualitative” components, as well as for the sleep
disturbances component, which may include items that are variable between
individuals and across time.

Consistency for the PSQI is lower than that reported by Domino et al. (1984), who
administered their questionnaire on two occasions separated by 10 weeks, and found
Pearson correlations for the seven factors of their questionnaire ranging from 0.68 to
0.79. However, this sample did not include patients with sleep or psychiatric
disorders.
Clinimetric Properties: Validity. The identification of “good” and “poor” sleepers
for research studies relies on subjective assessments of sleep quality, clinical
interviews, and polysomnographic studies. The PSQI provides a standardized,
quantitative measure of sleep quality that quickly identifies good and poor sleepers,
and compares favorably with the “gold standard” of clinical and laboratory
diagnosis. In the current study, good and poor sleepers consisted of healthy control
subjects and depressed or sleepdisordered patients. A global PSQI score > 5
provided a sensitive and specific measure of poor sleep quality, relative to clinical
and laboratory measures. Age and sex did not strongly correlate with PSQI
component scores, but they were significant covariates for the global score. Given
the differences in mean age and sex ratio between groups (with good sleepers being
older than poor sleepers), the current results are likely to underestimate the PSQI’s
ability to identify good and poor sleepers.
Distinct component score profiles emerged for controls, DOES patients, and
DIMS/depressed patients. The PSQI did not differentiate DIMS and depressed
patients. This is not surprising, since the sleep disturbance of depressives is most
often a sleep onset and maintenance insomnia. In fact, the current study lends
validity to the classification of depressive sleep disturbance as a DIMS in the ASDC
classification.
The PSQI is primarily intended to measure sleep quality and to identify good and
bad sleepers, not to provide accurate clinical diagnoses. Nevertheless, responses to
specific questions can point the clinician toward areas for further investigation. This
is particularly true for the “sleep disturbances”component, which may guide clinical
evaluations for specific patients, even though mean scores do not discriminate
between groups. Furthermore, a PSQI global score > 5 indicates that a subject is
having severe difficulties in at least two areas, or moderate difficulties in more than
three areas. The global score is therefore “transparent,” i.e., it conveys information
about the severity of the subject’s problem, and the number of problems present,
through a single simple measure (Feinstein, 1987).
A number of other studies have also validated their sleep questionnaires by
comparing different subject populations. For example, Domino et al. (1984)
validated their scale by administering it to patients with and without complaints of
sleep disturbance at a family physician’s office, and to patients with and without
depressive symptoms at a community mental health center. In each case, the patient
groups differed significantly on anumber of the questionnaire’s scales. Beutler et al.
(1978) used a stepwise discriminant function analysis to identify self-proclaimed
insomniacs and controls. Their discriminant model, including subject age and three
of the eight factors in their questionnaire, correctly identified 93% (86/92) of
subjects. A separate discriminant analysis correctly identified 86% (37/43) of
insomniacs who used or did not use medications. Mendelson et al. (1984, 1986)

206
found statistically significant differences on self-report sleep questionnaires for
insomniacs versus controls. McGhie (1966) compared depressed and nondepressed
psychiatric patients, and found no differences on the total sleep disturbance scale of
his questionnaire. Finally, Webb et al. (1976) found differences between “good”and
“bad” sleep episodes, and between high school and elderly subjects, with their
postsleep inventory. Like these previous questionnaires, the PSQI separates patients
with different diagnoses, but differs in that control subjects and patients in this study
were diagnosed according to current research and clinical criteria. In addition,
except for Beutler et al. (1978), previous reports have not indicated the sensitivity
and specificity of their scales.
A number of other studies have reported on the use of polysomnography to
validate subjective sleep reports (e.g., Lewis, 1969; Monroe, 1969; Baekeland and
Hoy, 1971; Bixler et al., 1973; Frankel et al., 1973, 1976; Carskadon et al., 1976;
Mendelson et al., 1984, 1986; Hoch et al., 1987). Several consistent findings emerge
from these reports. First, subjects with and without insomnia are not “accurate” in
their subjective report of variables such as sleep latency, sleep duration, and number
of arousals. However, while control subjects tend to overestimate their ability to
sleep, insomniacs tend to underestimate it, perhaps because they misperceive the
experience of being asleep (W. Mendelson, personal communication, April IS,
1988). Second, while subjective estimates and objective measures of sleep differ in
actual amount, they are often strongly and positively correlated. Finally, postsleep
questionnaires yield more accurate subjective estimates that are more strongly
correlated with polysomnographic findings. The current findings using the PSQI did
not replicate these general findings, as PSQI responses were not found to correlate
with polysomnographic measures. It is not surprising that subjects differed in
subjective and polysomnographic variables, since the PSQl asks for a global
estimate spanning I month, and is not sensitive to daily variability.
Applications. The PSQl’s simplicity and its ability to identify different groups of
patients suggest several clinical and research applications in psychiatry and general
medical settings. Most fundamentally, it may be used as a simple screening measure
to identify cases and controls, or “good” and “poor” sleepers. In a general clinical
setting, the PSQI could be used to screen patients for the presence of significant sleep
disturbance. In psychiatric settings, the PSQI may identify patients who are likely to
have a sleep disturbance concomitant with their psychiatric symptoms. In addition,
it may direct the clinican to specific areas of dysfunction that require further
investigation. The PSQI could also be used in clinical research and epidemiological
studies to identify groups that differ in the quality of their sleep.
The PSQI may also have several longitudinal applications in clinical practice and
research. For example, it could be used to examine the course and natural history of
sleep/wake disorders. It could also be used to monitor the progression of sleep
disturbances and their interaction with other symptoms during the course of
psychiatric illnesses such as depression. Rodin et al. (1988) recently published one of
the few studies to examine the interaction between depressive symptoms and sleep
disturbance longitudinally. The PSQI may be helpful in future studies of this type,
providing more detailed information about types and severity of sleep disturbances

over time. Further, the PSQl could be useful in studying the relation between sleep
quality and other variables, such as age, gender, health status, medical and
psychiatric conditions, and performance on other psychological variables. Finally,
the PSQI could be used to examine the longitudinal effects of specific therapeutic
interventions for psychiatric disorders or sleep disorders. For example, sleep quality
could be monitored during maintenance treatment of depression with medications or
psychotherapy. Used in this way, the PSQI might also detect relapses heralded by the
onset or reemergence of sleep disturbance.
Acknowledgment. This research was supported in part by NIMH grants MH-37869
(C.F.R.), MH-00295 (C.F.R.), MH-40023 (C.F.R.), MH-30915 (D.J.K.), and AC-06836
(T.H.M and C.F.R.).
References
American Psychiatric Association. DSM-III: Diagnostic and Statistical Manual of Mental
Disorders. 3rd ed. Washington, DC: APA, 1980.
Association of Sleep Disorders Centers. “Diagnostic Classification of Sleep and Arousal
Disorders.” 1st ed. (Prepared by the Sleep Disorders Classification Committee, H.P.
Roffwarg, Chairman) Sleep, Vol. 2, No. I, 1979.
Baekeland, F., and Hoy, P. Reported vs. recorded sleep characteristics. Archives of General
Psychiatry, 24:548-55 I, I97 I.
Beutler, L.E.; Thornby, J.I.; and Karacan, I. Psychological variables in the diagnosis of
insomnia. In: Williams, R.L., and Karacan, I., eds. Sleep Disorders: Diagnosis and
Treatment. New York: John Wiley & Sons, 1978. pp. 61-100.
Bixler, E.O.; Kales, A.; Soldatos, C.R.; Kales, J.D.; and Healey, S. Prevalence of sleep
disorders in the Los Angeles Metropolitan area. American Journal of Psychiatry, 136: I257-
1262, 1979.
Carskadon, M.A.; Dement, W.C.; Mitler, M.M.; Guilleminault, C.; Zarcone, V.P.; and
Spiegel, R. Self-reports versus sleep laboratory findings in 122 drug-free subjects with
complaints of chronic insomnia. Archives of General Psychiatry, 133: 1382-l 388, 1976.
Coleman, R.M. Diagnosis, treatment, and follow-up of about 8,000 sleep/wake disorder
patients. In: Guilleminault, C., and Lugaresi, E., eds. Sleep/ Wake Disorders: Natural History,
Epidemiology, and Long-Term Evolution. New York: Raven Press, 1983. pp. 87-97.
Consensus conference: Drugs and insomnia, the use of medications to promote sleep.
Journal of the American Medical Association, 25 I 124 I O-24 14, 1984.
Cronbach, L.J. Coefficient alpha and the internal structure of tests. Psychometrika, 16:297-
334,195l.
Domino, G.; Blair, G.; and Bridges, A. Subjective assessment of sleep by sleep
questionnaire. Perceptual and Motor Skills, 59: I63- 170, 1984.
Ellis, B.W.; Johns, M.W.; Lancaster, R.; Raptopoulos, P.; Angelopoulos, N.; and Priest,
R.G. Short report: The St. Mary’s Hospital Sleep Questionnaire: A study of reliability. Sleep,
4:93-97, I98 I.
Endicott, J., and Spitzer, R.L. A diagnostic interview: Schedule for Affective Disorders and
Schizophrenia. Archives of General Psychiatry, 35:837-844, 1978.
Feinstein, A.R. Clinimetrics. New Haven, CT: Yale University Press, 1987.
Frankel, B.L.; Buchbinder, R.; and Snyder, F. Ineffectiveness of electrosleep in chronic
primary insomnia. Archives of General Psychiatry, 29:563-568, 1973.
Frankel, B.L.; Coursey, R.D.; Buchbinder, R.; and Snyder, F. Recorded and reported sleep
in chronic primary insomnia. Archives of General Psychiatry, 33:6 15-623, 1976.
Hamilton, M. A rating scale for depression. Journal of Neurology, Neurosurgery, and
Psychiatry, 23~56-62, 1960.

208
Hoch, C.C.; Reyno1dsC.F. 111; Kupfer, D.J.; Berman, S.R.; Houck, P.R.;and Stack, J.A.
Empirical note: Self-report versus recorded sleep in healthy seniors. Psychophysiology,
24:293-299, 1987.
Jacobs, E.A.; Reynolds, C.F. Ill; Kupfer, D.J.; Lovin, P.A.; and Ehrenpreis, A.B. The role
of polysomnography in the differential diagnosis of chronic insomnia. American Journal of
Psychiatry, 145346-349, 1988.
Johnson, L.C., and Spinweber, C.L. Quality of sleep and performance in the Navy: A
longitudinal study of good and poor sleepers. In: Guilleminault, C., and Lugaresi, E., eds.
Sleep / Wake Disorders: Natural History, Epidemiology. and Long- Term Evolution. New
York: Raven Press, 1983. pp. 13-28.
Karacan, 1.; Thornby, J.I.; Anch, M.; Holzler, C.E.; Warheit, G.J.; Schivab, J.J.; and
Williams, R.L. Prevalance of sleep disturbances in a primarily urban Florida county. Social
Science & Medicine, 10:239-244, 1976.
Karacan, I.; Thornby, J.I.; and Williams, R.L. Sleep disturbance: A community survey. In:
Guilleminault, C., and Lugaresi, E., eds. Sleep/ Wake Disorders: Natural History,
Epidemiology, and Lang-Term Evolution. New York: Raven Press, 1983. pp. 37-60.
Kripke, D.F.; Simons, R.N.; Garfinkiel, L.; and Hammond, E.C. Short and long sleep and
sleeping pills: Is increased mortality associated? Archives of General Psychiatry, 36: 103-I 16,
1979.
Lewis, S.A. Subjective estimates of sleep: An EEG evaluation. British Journal of
Psychology, 60:203-209, 1969.
Lugaresi, E.; Cirignotta, F.; Zucconi, M.; Mandini, S.; Lenzi, P.L.; and Coccagna, G. Good
and poor sleepers: An epidemiological survey of the San Marion population. In:
Guilleminault, C., and Lugaresi, E., eds. Sleep/ Wake Disorders: Natural History,
Epidemiology, and Lang- Term Evolution. New York: Raven Press, 1983. pp. I - 12.
McGhie, A. The subjective assessment of sleep patterns in psychiatric illness. British
Journal of Medical Psychology, 39~22 I-230, 1966.
McGhie, A., and Russell, S.M. The subjective assessment of normal sleep patterns. Journal
of Mental Science, 108642-654, 1962.
Mellinger, G.D.; Balter, M.B.; and Uhlenhuth, E.H. Insomnia and its treatment: Prevalence
and correlates. Archives of General Psychiatry, 421225232, 1985.
Mendelson, W.B.; Garnett, D.; Gillin, J.C.; and Weingartner, H. The experience of
insomnia and daytime and nighttime functioning. Ps_ychiatry Research, 12:235-250, 1984.
Mendelson, W.B.; James, S.P.; Garnett, D.; Sack, D.A.; and Rosenthal, N.E. A
psychophysiological study of insomnia. Psychiatry Research. 19:267-284, 1986.
Monroe, L.J. Psychological and physiological differences between good and poor sleepers.
Journal of Abnormal Psychology, 72:255-264, 1969.
Parrott, A.C., and Hindmarch, I. Factor analysis of a sleep evaluation questionnaire.
Psychological Medicine, 81325-329, 1978.
Parrott, A.C., and Hindmarch, I. The Leeds Sleep Evaluation Questionnaire in
psychopharmacological investigations-A review. Psychopharmacology. 7 I : 173-l 79, 1980.
Rechtschaffen, A., and Kales, A., eds.; Berger, R.J.; Dement, W.C.; Jacobson, A.; Johnson,
L.C.; Jouvet, M.; Monroe, L.J.; Oswald, I.; Roffwarg, H.P.; Roth, B.; and Walter, R.D. A
Manual of Standardized Terminology, Techniques and Scoring System for Sleep Stages of
Human Subjects. Washington, DC: Superintendent of Documents, U.S. Government
Printing Office, 1968.
Rodin, J.; McAvay, G.; and Timko, C. A longitudinal study of depressed mood and sleep
disturbances in elderly adults. Journal of Gerontology: Psychological Sciences, 43145-53,
1988.
Samuel, J.G. Sleep disturbances in depressed patients: Objective and subjective measures.
British Journal of Psychiatry, I IO:7 I l-7 19, 1964.
Spitzer, R.L.; Endicott, J.; and Robins, E. Research Diagnostic Criteria: Rationale and
reliability. Archives of General Psychiatry, 35:773-782, 1978.
Webb, W.B.; Bonnet, M.; and Blume, G. A post-sleep inventory. Perceptual and Motor
Skills, 43:987-993, 1976.

209
Welstein, L.; Dement, W.C.; Redington, D.; and Guilleminault, C. Insomnia in the San
Francisco Bay Area: A telephone survey. Sleep/ Wake Disorders: Natural History,
Epidemiology, and Long-Term Evolution. New York: Raven Press, 1983. pp. 73-85.
Appendix. Pittsburgh Sleep Quality Index (PSQI)
Name ID # Date Age ___
Instructions:
The following questions relate to your usual sleep habits during the past month only. Your answers
should indicate the most accurate reply for the majority of days and nights in the past month.
Please answer all questions.
1. During the past month, when have you usually gone to bed at night?
USUAL BED TIME
2. During the past month, how long (in minutes) has it usually take you to fall asleep each night?
NUMBER OF MINUTES
3. During the past month, when have you usually gotten up in the morning?
USUAL GETTING UP TIME
4. During the past month, how many hours of actual sleep did you get at night? (This may be different
than the number of hours you spend in bed.)
HOURS OF SLEEP PER NIGHT
For each of the remaining questions, check the one best response. Please answer a// questions.
5. During the past month, how often have you had trouble sleeping because you...
(a) Cannot get to sleep within 30 minutes
Not during the Less than Once or Three or more
past month _ once a week - twice a week _ times a week _
(b) Wake up in the middle of the night or early morning
Not during the Less than Once or Three or more
past month - once a week - twice a week _ times a week _
(c) Have to get up to use the bathroom
Not during the Less than Once or Three or more
past month - once a week - twice a week - times a week -
(d) Cannot breathe comfortably
Not during the Less than Once or Three or more
past month - once a week - twice a week - times a week _
(e) Cough or snore loudly
Not during the
past month -
(f) Feel too cold
Not during the
past month _
(g) Feel too hot
Not during the
past month _
(h) Had bad dreams
Not during the
past month _
(i) Have pain
Not during the
past month _
Less than
once a week -
Less than
once a week -
Less than Once or Three or more
once a week - twice a week - times a week _
Less than
once a week -
Less than
once a week -_
Once or Three or more
twice a week - times a week _
Once or Three or more
twice a week - times a week -
Once or Three or more
twice a week - times a week _
Once or Three or more
twice a week - times a week _

210
(j) Other reason(s), please describe
How often during the past month have you had trouble sleeping because of this?
Not during the Less than Once or Three or more
past month - once a week - twice a week _ times a week -
6. During the past month, how would you rate your sleep quality overall?
Very good ~
Fairly good ~
Fairly bad
Very bad ___
7. During the past month, how often have you taken medicine (prescribed or “over the counter”) to help
you sleep?
Not during the Less than Once or Three or more
past month - once a week ~ twice a week - times a week -
8. During the past month, how often have you had trouble staying awake while driving, eating meals, or
engaging in social activity?
Not during the Less than Once or Three or more
past month ~ once a week ~ twice a week - times a week _
9. During the past month, how much of a problem has it been for you to keep up enough enthusiasm to
get things done?
No problem at all
Only a very slight problem
Somewhat of a problem ~
A very big problem ~
10. Do you have a bed partner or roommate?
No bed partner or roommate
Partner/roommate in other room
Partner in same room, but not same bed
Partner in same bed
If you have a roommate or bed partner, ask him/her how often in the past month you have had...
(a) Loud snoring
Not during the Less than Once or Three or more
past month ~ once a week ~ twice a week - times a week ___
(b) Long pauses between breaths while asleep
Not during the Less than Once or Three or more
past month - once a week ~ twice a week - times a week _
(c) Legs twitching or jerking while you sleep
Not during the Less than Once or Three or more
past month ~ once a week _ twice a week ~ times a week __
(d) Episodes of disorientation or confusion during sleep
Not during the Less than Once or Three or more
past month - once a week - twice a week - times a week __
(e) Other restlessness while you sleep: please describe
Not during the Less than Once or Three or more
past month ~ once a week - twice a week - times a week -

211
Scoring Jnstructions for the Pittsburgh Sleep Quality Index
The Pittsburgh Sleep Quality Index (PSQI) contains 19 self-rated questions and 5 questions rated by the
bed partner or roommate (if one is available). Only self-rated questions are included in the scoring. The
19 self-rated items are combined to form seven “component” scores, each of which has a range of O-3
points. In all cases, a score of “0” indicates no difficulty, while a score of “3” indicates severe difficulty.
The seven component scores are then added to yield one “global” score, with a range of O-21 points,
“0” indicating no difficulty and “21” indicating severe difficulties in all areas.
Scoring proceeds as follows:
Component 1: Subjective sleep quality
Examine question #6, and assign scores as follows:
Response Component 1 score
“Very good” 0
“Fairly good” 1
“Fairly bad” 2
‘Very bad” 3
Component 1 score: -
Component 2: Sleep latency
1. Examine question #2, and assign scores as follows:
Response Score
I 15 minutes 0
16-30 minutes 1
31-60 minutes 2
> 60 minutes 3
Question #2 score:
2. Examine question #5a, and assign scores as follows:
Response Score
Not during the past month 0
Less than once a week 1
Once or twice a week 2
Three or more times a week 3
Question #5a score:
3. Add #2 score and #5a score
Sum of #2 and #5a: ~
4. Assign component 2 score as follows:
Sum of #2 and #5a Component 2 score
0 0
l-2 1
3-4 2
5-6 3
Component 2 score: ___
Component 3: Sleep duration
Examine question #4, and assign scores as follows:
Response Component 3 score
> 7 hours 0
6-7 hours 1
5-6 hours 2
< 5 hours 3
Component 3 score: ~

212
Component 4: Habitual sleep efficiency
(1) Write the number of hours slept (question # 4) here: ~
(2) Calculate the number of hours spent in bed:
Getting up time (question # 3): ~
- Bedtime (question # 1):
Number of hours spent in bed: ~
(3) Calculate habiiual sleep efficiency as foflows:
(Number of hours slept/Number of hours spent in bed) X 100 = Habitual sleep efficiency (%)
(/)X100= %
(4) Assign component 4 score as follows:
Habitual sleep efficiency %
> 65%
7564%
65-74%
< 65%
Component 5: Sleep disturbances
Component 4 score
Component 4 score:
(1) Examine questions # 5b-5j, and assign scores for each question as follows:
Response Score
Not during the past month 0
Less than once a week 1
Once or twice a week 2
Three or more times a week 3
#5b score ~
c score
d score ~
e score
f score ~
g score ~
h score
i score
j score ~
(2) Add the scores for questions # 5b-5j:
Sum of # 5b-5j:
(3) Assign component 5 score as follows:
Sum of # 5b-5j Component 5 score
0 0
l-9 1
10-16 2
19-27 3
Component 5 score: ___
Component 6: Use of sleeping medication
Examine question # 7 and assign scores as follows:
Response Component 6 score
Not during the past month 0
Less than once a week 1
Once or twice a week 2
Three or more times a week 3
Component 6 score: ~

213
Component 7: Daytime dystunction
(1) Examine question # 8, and assign scores as foflows:
Response Score
Never 0
Once or twice 1
Once or twice each week 2
Three or more times each week 3
Question tt 8 score: ~
(2) Examine question t 9, and assign scores as follows:
Response Score
No problem at aft 0
Only a very slight problem 1
Somewhat of a probfem 2
A very big problem 3
Question tt 9 score: -
(3) Add the scores for question # 8 and ft 9:
Sum of tt8 and #9: ~
(4) Assign component 7 score as follows:
Sumoftf8andft9 Component 7 score
0 0
l-2 1
3-4 2
5-6 3
Global PSGI Score
Add the seven component scores together:
Component 7 score: ___
Global PSGI Score: ~