Arthritis & Rheumatism (Arthritis Care & Research)

Vol. 49, No. 5S, October 15, 2003, pp S134 –S146

DOI 10.1002/art.11410
© 2003, American College of Rheumatology
MEASURES OF PSYCHOLOGICAL STATUS AND WELL-BEING

Measures of Depression and Depressive

Symptoms
The Beck Depression Inventory (BDI), Center for Epidemiological Studies-
Depression Scale (CES-D), Geriatric Depression Scale (GDS), Hospital Anxiety
and Depression Scale (HADS), and Primary Care Evaluation of Mental Disorders-
Mood Module (PRIME-MD)

Karen L. Smarr

BECK DEPRESSION INVENTORY (BDI) cognitive and affective items from the BDI-II to
assess depression in individuals with biomedical
General Description or substance abuse problems (5). The BDI
Purpose. To measure depression symptoms and FastScreen excludes the somatic items from BDI-II.
severity in persons age 13 and older. The timeframe on the BDI FastScreen is the same
as BDI-II.
Content. Based on clinical observations and The BDI has been translated into several
patient description, the BDI contains items that languages, including Spanish, Chinese, Dutch,
reflect the cognitive, affective, somatic, and Finnish, French (Canadian), German, Korean,
vegetative symptoms of depression (1,2). Polish, Swedish, and Turkish (6).
Developer/contact information. Aaron T. Beck, Number of items in scale. There are 21 items in
PhD, Center for Cognitive Therapy, Philadelphia, the BDI-IA and BDI-II; and 7-items in the BDI
PA. FastScreen.
Versions. The revised BDI (BDI-IA) (3), replaced Subscales. None typically reported. The BDI-IA
the original version (2). The BDI-IA is similar to manual discusses the cognitive-affective (Items 1–
the original, except timeframe extends “over the 13) and the somatic-performance (Items 14 –21)
past week, including today” and some items were subscales that discriminate between psychiatric,
reworded to avoid double negative statements. The medical, and normal samples (2). Factors analysis
BDI-II contains substantial revision of the original of the BDI-II revealed two intercorrelated factors,
and revised BDI so that the assessment of somatic-affective and cognitive dimensions (3).
symptoms corresponds to the DSM-IV criteria (4).
BDI-II timeframe extends for 2 weeks to correspond Populations. Developmental/target. BDI-IA:
with the DSM-IV criteria for major depressive developed and validated using psychiatric and
disorder. normal populations. Beck and colleagues (3)
BDI FastScreen for Medical Patients (formerly studied outpatient samples that included persons
known as BDI-Primary Care [BDI-PC]) contains 7 with severe psychiatric diagnoses, depressive
disorders, substance abuse, and college students.
Supported by the National Institute on Disability and BDI-II validated using college students, adult
Rehabilitation Research (H133B30039) and the Medical Re- psychiatric outpatients, and adolescent psychiatric
search Service of the Department of Veterans Affairs.
Karen L. Smarr, PhD: University of Missouri-Columbia, outpatients (4).
School of Medicine and the Harry S Truman Memorial BDI-FastScreen validated using general
Veterans’ Hospital. medical inpatients referred for psychiatric
Address correspondence to Karen L. Smarr, PhD, Harry S
Truman Memorial Veterans’ Hospital, Behavioral Health consultation and outpatients seen by family
Service Line, Columbia, MO 65201. E-mail: practice, pediatrics, and internal medicine (5).
[email protected].
Submitted for publication April 23, 2003; accepted April
24, 2003.
Other uses. Since being revised in 1972, the BDI
has been widely accepted and used in psychology

S134
Depression S135

and psychiatry for assessing the intensity of of depression.The following guidelines have been
depression in psychiatric and normal populations. suggested to interpret the revised BDI (the BDI-IA)
Studies have been conducted in a variety of (3). With the normal population, a BDI-IA score of
settings using medical populations (e.g., ⱖ15 may indicate possible depression and
Parkinson’s disease, human immunodeficiency warrants an additional clinical evaluation as
virus, oncology), persons with disabilities, (e.g., confirmation. Minimal range 0 –9; Mild depression
arthritis, spinal cord injury, amputation), veterans, 10 –16; Moderate depression 17–29; Severe
students, older adults, adolescents, and many depression 30 – 63.
populations with psychiatric diagnoses (e.g., eating The following guidelines have been suggested
disorders, addictions, anxiety disorders). to interpret the BDI-II (4), Minimal range 0 –13;
Mild depression 14 –19; Moderate depression 20 –
WHO ICF Components. Participation restriction. 28; Severe depression 29 – 63.
The following guidelines have been suggested
Administration to interpret the BDI FastScreen for Medical
Method. Paper and pencil self-report in group or Patients (5). Minimal 0 –3; Mild depression 4 – 8;
individual format; self or oral administration. Moderate depression 9 –12; Severe depression 13–
21.
Training. Minimal training required for
paraprofessionals or professionals to administer. A Method of scoring. Sum the severity ratings of
clinician needs to interpret the revised BDI score each depression item. Use the highest response
by paying particular attention to items endorsing when an item has more than one severity rating.
self harm or feelings of helplessness, such as Special instructions: BDI-IA: If examinee is
suicide ideation (item 9) and consciously trying to lose weight, then Item 19 is
pessimism/hopelessness (item 2). not added to total score. BDI-II: For diagnostic
purposes, Item 16 (sleep patterns changes) and
Time to administer/complete. Item 18 (appetite changes) contain 7-point ratings
Self-administration: 5-10 minutes; Oral to note increases or decreases in behavior.
administration: 15 minutes.
Time to score. 5 minutes.
Equipment needed. Pencil or pen to indicate
response. Training to score. Minimal training, 5–10
minutes.
Cost/availability. Contact The Psychological
Corporation to purchase the BDI, BDI-II, or the BDI Training to interpret. Minimal training to
FastScreen for Medical Patients manuals and interpret, yet due to the suicide risk with
instrument. Computer software is available from depression, a health professional should interpret
Psychological Corporation for on-screen the BDI-IA to provide appropriate referrals and
administration, for use with paper and pencil possibly psychotherapeutic interventions for at-risk
administration, or for input of data from a desktop individuals.
scanner. The computer program may be used to
administer a single questionnaire or to integrate Norms available. Means and standard deviations
the results of sequential administrations. The appear in the manuals for samples used to validate
Psychological Corporation, 555 Academic Court, the instrument.
San Antonio, TX 78204; Website:
www.psychcorp.com. Items may be seen in
McDowell and Newell (7). Psychometric Information
Reliability. Internal consistency. Beck and Steer
Scoring (3) report that Cronbach’s coefficient alphas for the
Responses. Scale. 4-point scale indicates degree revised BDI’s normative-psychiatric samples range
of severity; items are rated from 0 (not at all) to 3 from 0.79 to 0.90. These coefficients are consistent
(extreme form of each symptom). with estimates of coefficient alpha reported in a
psychiatric sample (0.86) and in a non-psychiatric
Score range. BDI: 0 – 63; BDI-II: 0 – 63; BDI sample (0.81; 8). The BDI-II has higher internal
FastScreen: 0 –21. consistency than the BDI-IA: Cronbach’s alpha
reported as 0.92 for outpatients and 0.93 for
Interpretation of scores. No arbitrary cut-off college students. Coefficient alphas for BDI
score for all purposes to classify different degrees FastSceen ranged from 0.85 to 0.89.
S136 Smarr

Test-retest. Beck et al’s (8) review of BDI-IA References

studies reported correlations between pre- and 1. (Original) Beck AT, Ward CH, Mendelson M, Mock J,
posttests, for varying time intervals, that ranged Erbaugh J. An inventory for measuring depression.
from 0.48 to 0.86 for psychiatric patients and from Arch Gen Psychiatry 1961;4:561–71.
0.60 to 0.90 for non-psychiatric patients. For 2. Beck AT, Rush AJ, Shaw BF, Emery G. Cognitive
college students, test-retest correlations ranged therapy of depression. New York: Guilford Press;
1979.
from 0.64 to 0.90; BDI-II test-retest (administered
3. Beck AT, Steer RA. Manual for the Beck Depression
1-week apart) correlation was 0.93. Inventory, 1993 edition. San Antonio (TX): The
Psychological Corporation; 1987.
Validity. Content. According to the manual (3), 4. Beck AT, Steer RA, Brown GK. Beck Depression
BDI-IA items reflect 6 of 9 DSM-II criteria well. Inventory-Second Edition Manual. San Antonio (TX):
The BDI-II revision improved content validity by The Psychological Corporation; 1996.
rewording and adding items to assess DSM-IV 5. Beck AT, Steer RA, Brown GK. BDI-FastScreen for
criteria for depression. Medical Patients Manual. San Antonio (TX): The
Psychological Corporation; 2000.
6. Naughton MJ, Wiklund I. A critical review of
Construct. As theorized, the BDI-IA and BDI-II
dimension-specific measures of health-related quality
are positively correlated with hopelessness of life in cross-cultural research. Qual Life Res 1993;2:
construct in normative samples. In a factor analysis 397– 432.
of the BDI responses of patients and non-patients, 7. McDowell I, Newell CI. Measuring health: a guide to
Beck and colleagues (8) found that 3 factors rating scales and questionnaires. New York: Oxford
(cognitive-affective, performance, and somatic) University Press; 1996.
were consistently identified across diagnostic 8. Beck AT, Steer RA, Garbin M. Psychometric properties
of the Beck Depression Inventory: twenty-five years of
groups. Factor analysis of the BDI-II yielded 2
evaluation. Clin Psychol Rev 1988;8:77–100.
factors (somatic-affective and cognitive factors) (4). 9. Callahan LF, Kaplan MR, Pincus T. The Beck
Depression Inventory, Center for Epidemiological
Criterion. Beck and colleagues (8) reported a Studies Depression Scale (CES-D), and General Well-
mean correlation of 0.72 between BDI-IA and Being Schedule Depression subscale in rheumatoid
clinical depression ratings in psychiatric patients arthritis. Arthritis Care Res 1991;4:3–11.
and 0.60 in a nonpsychiatric sample. In normative
samples, correlations between BDI-IA and
Hamilton Rating Scale for Depression (HRSD)
Additional References
ranged from 0.40 in single episode major Beck AT, Guth D, Steer RA, Ball R. Screening for major
depression disorders in medical inpatients with the
depression to 0.87 in alcoholics. The BDI-IA was
Beck Depression Inventory for primary care. Behav
significantly related to the depression subscale of Res Ther 1997;35:785–91.
the Symptom Checklist-90-Revised (0.76); BDI-II Beck AT, Steer RA, Ball R, Ranieri WF. Comparison of the
and HRSD were positively correlated (0.71) (4). Beck Depression Inventories-1A and -II in psychiatric
outpatients. J Pers Assess 1996;67:588 –97.
Steer RA, Cavalieri TA, Leonard DM, Beck AT. Use of
Sensitivity/responsiveness to change. The BDI is
the Beck Depression Inventory for Primary Care to
less likely to overestimate changes due to screen for major depression disorders. Gen Hosp
psychotherapy or pharmacologic interventions than Psychiatry 1999;21:106 –11.
the HRSD. Steer RA, Rissmiller DJ, Beck AT. Use of the Beck
Depression Inventory-II with depressed geriatric
inpatients. Behav Res Ther 2000;38:311– 8.
Tanaka-Matsumi J, Kameoka VA. Reliabilities and
Comments and Critique concurrent validities of popular self-report measures
The BDI has been criticized for having items of depression, anxiety, and social desirability. J
Consult Clin Psychol 1986;54:328 –33.
that are confounded by the physical sequelae
associated with physical disability, such as
arthritis; items related to physical appearance, CENTER FOR EPIDEMIOLOGICAL
fatigue, ability to work, weight loss, and physical
STUDIES-DEPRESSION SCALE (CES-D)
complaints (9). The cost of the test materials may
be prohibitive since less expensive public domain General Description
assessments are readily available. The manual Purpose. To measure current level of depressive
suggests using the BDI cautiously as a screening symptomatology in a general population. The CES-
tool for clinical depression. D was developed for research purposes and is used
Depression S137

as a screening tool to identity persons at risk for Equipment needed. When self-administered,
clinical depression (1). need a pencil or pen to complete.

Content. Items assess perceived mood and level Cost/availability. The CES-D is available in
of functioning during the past week. Four factors original article by Radloff (1) and is available from
are represented: depressed affect, positive affect, the National Institutes of Health, Epidemiology
somatic problems and retarded activity, and Branch, 5600 Fishers Lane, Rockville, MD 20857;
interpersonal relationship problems. Items do not available at www.chcr.brown.edu/pcoc/cesdscale.pdf
assess the diagnostic criteria of appetite, and www.psychiatry.uchc.edu/screening/CES-D/.
anhedonia, psychomotor agitation or retardation, There is no cost to use the CES-D.
guilt, or suicidality.
Scoring
Developer/contact information. Lenore Sawyer Responses. Scale. 4-point scale, where 0 ⫽
Radloff, National Institute of Mental Health, rarely or none of the time (less than 1 day), 1 ⫽
Rockville, MD. some or a little of the time (1–2 days), 2 ⫽
occasionally or a moderate amount of time (3– 4
Versions. Translated into Spanish, Chinese, days), 3 ⫽ most or all the time (5–7 days).
Dutch, Korean, Russian, German, and French. The
original 20-item version has been shortened to a Score range. The range is 0 – 60.
10-item version for older adults (2) and to 5-item
version (3). There is also a modified version Interpretation of scores. A higher score reflects
available for children (Center for Epidemiological greater symptoms of depression, weighted by
Studies-Depression Scale [CES-] for Children [CES- frequency of occurrence in past week. CES-D ⱖ16
DC] (4). is typically employed as a cut-off for clinical
depression and usually warrants a referral for a
Number of items in scale. There are 20 items. more thorough evaluation.
Turk and Okifuji (6) recommend a cut-off
Subscales. A total score is obtained, no score of 19 for detecting depressive disorder in
subscales. chronic pain patients. Blalock et al (7) identified 4
arthritis-related items and suggested a modified
Populations. Developmental/target. scoring approach. Callahan et al (8) discussed
Epidemiology studies using a general population. additional scoring issues in rheumatic disease.
Other uses. Widely used and validated in many Method of scoring. Easily hand scored. Items are
populations including rheumatoid arthritis, summed to obtain a total score, using the 0 (rarely
fibromyalgia, and other medical cohorts (stroke, or none of the time) to 3 (most or all the time)
multiple sclerosis, oncology); adolescents; women; scores for individual items. Four items (4, 8, 12,
African Americans; primary care; the elderly; 16) are worded in a positive direction to reduce a
persons of Korean, Puerto Rican, Spanish, and tendency towards response bias; these items are
Chinese decent; American Indians; and in clinical reverse scored.
and psychiatric populations. Reliability and
validity available for African American, Asian Time to score. Less than 10 minutes. Can be
American, French, Greek, Hispanic, Japanese, and scored during administration.
Yugoslavian populations (5).
Training to score. Minimal training time to
WHO ICF Components. Participation restriction. score, ⱕ10 minutes.

Administration Training to interpret. ⱕ10 minutes.

Method. Easily self-administered or
administered by interviewer. Can be administered Norms available. Reference scores available at
in-person, written, or interview format, by Sayette and Johnson (9), Eaton and Kessler (10),
telephone interview, or by mail. Murrell et al (11), Golding et al (12), and Vera et al
(13).
Training. Minimal.
Psychometric Information
Time to administer/complete. Approximately 10 Reliability. Internal consistency. High internal
minutes. consistency. Coefficient alphas range from 0.85 in
S138 Smarr

the general population to 0.90 in a psychiatric scoring system (yes/no) (15). It is not intended as a
patient population. diagnostic tool (16), to discriminate among
depression subtypes (major depressive disorder
Test-retest. The CES-D measures “current” level versus dysthymic disorder; bipolar versus
of symptomatology and is expected to vary over unipolar), or to distinguish between a primary or
time. In the original sample, test-retest correlations secondary depression (17). The CES-D is
were in the moderate range falling between 0.45 appropriate as a screening tool to identify
and 0.70, as expected if the scale is sensitive to depressive disorders in clinical and research
current depressive state; stronger test-retest settings and as the initial element of a two-pronged
correlations were identified with shorter depression screening procedure in the general
administration time intervals. community, medical community, or primary care.
The CES-D provides a rough indicator of clinical
Validity. Content. Items were selected from depression and is modestly related to depressive
longer previously used and validated scales disorders.
considered to be representative of clinical The high correlation between CES-D and trait
symptoms of depression. anxiety indicates that CES-D measures depression
as well as anxiety, a conceptually related
Criterion. The CES-D adequately correlates with construct. Based on the validity studies, the CES-D
other valid self-report depressions scales to may not be specific for depression, but may be a
provide concurrent validity. In the original sample, measure of general distress.
CES-D correlations with depression measures (e.g., A CES-D cut-off score of 16 seems appropriate
Lubin, Bradburn Negative Affect) ranged from 0.51 in most populations, especially when the goal is to
to 0.61; moderate correlation (0.49) was found identify individuals at high risk for major
between CES-D and clinical interview ratings of depressive disorder, accepting some false positives.
depression. Using a cut-off of 16, CES-D scores Slightly lowering the CES-D cut-off may be
were found to discriminate strongly between necessary to identify persons with dysthymic
psychiatric inpatients and general population disorder or minor depressive disorder.
samples (70% with scores ⱖ16 versus 21%). CES-D
scores and depression severity ratings by a nurse
References
clinician were moderately correlated (0.56); CES-D
was negatively correlated (⫺0.20) with Bradburn 1. (Original) Radloff LS. The CES-D scale: a self-report
depression scale for research in the general
Positive Affect and low correlations (0.19 – 0.26)
population. Appl Psychol Meas 1977;1:385– 401.
were found between CES-D scores and 2. Irwin M, Artin KH, Oxman M. Screening for
medications, disability days, social functioning, depression in the older adult: criterion validity of
and aggression. CES-D scores were moderately the 10-item Center for Epidemiological Studies
correlated with self-esteem (0.58) and state anxiety Depression Scale (CES-D). Arch Intern Med 1999;
(0.44) and highly correlated (0.71) with trait 159:1701– 4.
anxiety (14). 3. Shrout PE, Yager TJ. Reliability and validity of
screening scales: effect of reducing scale length.
Sensitivity/responsiveness to change. Sensitive J Clin Epidemiol 1989;42:69 –78.
to change since the test-retest changes have been 4. Fendrich M, Weissman M, Warner V. Screening for
found before and after treatment, as well as before depressive disorder in children and adolescents:
validating the Center for Epidemiologic Studies
and after a stressful life event.
Depression Scale for Children. Am J Epidemiol 1990;
131:538 –51.
5. Naughton MJ, Wiklund I. A critical review of
Comments and Critique dimension-specific measures of health-related quality
The CES-D has been extensively used and of life in cross-cultural research. Qual Life Res 1993;
studied, and is considered a reliable valid 2:397– 432.
instrument; widely recognized research tool. It can 6. Turk DC, Okifuji A. Detecting depression in chronic
be used to measure change in affective state and is pain patients: adequacy of self-reports. Behav Res
Ther 1994;32:9 –16.
an excellent choice to measure depression
7. Blalock SJ, DeVellis RF, Brown GK, Wallston
symptoms in research studies. It can be used in KA.Validity of the Center for Epidemiological
diverse settings and has been validated in Studies Depression Scale in arthritis populations.
numerous populations, allowing comparisons Arthritis Rheum 1989;32:991–7.
across studies. 8. Callahan LF, Kaplan MR, Pincus T. The Beck
Researchers have used shorter versions to Depression Inventory, Center for Epidemiological
examine alternate CES-D cutoffs and simplified Studies Depression Scale (CES-D), and General Well-
Depression S139

Being Schedule Depression Subscale in rheumatoid Content. Items represent characteristics of

arthritis. Arthritis Care Res 1991;4:3–11. depression in the elderly in the affective (e.g.,
9. Sayetta R, Johnson D. Basic data on depressive sadness, apathy, crying) and cognitive domains
symptomatology, United States, 1974 –75.
(e.g., thoughts of hopelessness, helplessness, guilt,
Washington (DC): United States Government Printing
Office, Public Health Services 1980. (DHEW [PHS]
worthlessness). The GDS contains no somatic
80-1666). concerns common in elderly (i.e., disturbances in
10. Eaton WW, Kessler LG. Rates of symptoms of energy level, appetite, sleep, sexual interest).
depression in a national sample. Am J Epidemiol
1981;114:528 –38. Developer/contact information. Jerome Yesavage,
11. Murrell SA, Himmelfarb S, Wright K. Prevalence of MD, Stanford University, Director Aging Clinical
depression and its correlates in older adults. Am J Research Center. Stanford, CA.
Epidemiol 1983;117:173– 85.
12. Golding JM, Aneshensel CS, Hough RL. Responses to Versions. Long version (30 items) (1,2). Short
depression scale items among Mexican-Americans
version (15 items) developed to decrease fatigue or
and non-Hispanic whites. J Clin Psychol 1991;47:61–
75. lack of focus seen in the elderly (3). GDS has been
13. Vera M, Alegria M, Freeman D, Robles RR, Rios R, adapted into Japanese, Portuguese, Italian, Spanish,
Rios CF. Depressive symptoms among Puerto Ricans: Chinese, Korean, French, Swedish, and other
island poor compared with residents of the New languages. (See availability on website).
York City area. Am J Epidemiol 1991;134:502–10.
14. Orme JG, Reis J, Herz EJ. Factorial and discriminant Number of items in scale. There are 30 items
validity of the Center for Epidemiological Studies (long form); 15 items (short form).
Depression (CES-D) Scale. J Clin Psychol 1986;42:
28 –33. Subscales. None.
15. Kohout FJ, Berkman LF, Evans DA, Cornoni-Huntley
J. Two shorter forms of the CES-D depression
symptoms index. J Aging Health 1993;5:179 –93.
Populations. Developmental/target. Normal
16. Radloff LS, Teri L. Use of the Center for community dwelling elderly and elders
Epidemiological Studies-Depression Scale with older hospitalized for depression.
adults. Clin Gerontol 1986;5:119 –136.
17. Radloff LS, Locke BZ. The community mental health Other uses. Validity studied extensively. Stiles
assessment survey and the CES-D Scale. In: and McGarrahan (4) reported that the GDS has
Weismann MM, Meyers JK, Ross CG, editors. been used successfully in community samples,
Community survey of psychiatric disorder. New psychiatric and medical patients, nursing home
Brunswick, (NJ): Rutgers University Press; 1985. p.
residents (cognitively impaired and intact),
177– 89.
geriatric samples, and young adults. The GDS has
been used with medical patients, (i.e., stroke,
rheumatology, and cancer), African Americans,
Additional References
Mexican Americans, and other non-whites.
DeForge BR, Sobal J. Self-report depression scales in the
elderly: the relationship between the CES-D and the
WHO ICF Components. Participation restriction.
Zung. Int J Psychiatry Med 1988;18:325– 8.
Gerety MB, Williams JW Jr, Mulrow CD, Cornell JE,
Kadri AA, Rosenberg J, et al. Performance of case-
finding tools for depression in the nursing home: Administration
influence of clinical and functional characteristics
Method. Designed as paper and pencil, self-
and selection of optimal threshold scores. J Am
Geriatr Soc 1994;42:1103–9.
administered questionnaire. Oral assistance and/or
Roberts RE. Reliability of the CES-D Scale in different interview can be utilized; however, it has been
ethnic contexts. Psychiatry Res 1980;2:125– 43. suggested that the same format be used for
repeated administrations for patients or within
subject groups because different administration
GERIATRIC DEPRESSION SCALE (GDS) formats can produce variable results. Oral
administration may be advisable in some
General Description situations, particularly for individuals who have
Purpose. Developed as a self-rating, screening cognitive impairments.
tool to measure depressive symptoms in older
adults. Designed to identify depression in the Training. None.
elderly by distinguishing symptoms of depression
and dementia. Time to administer/complete. 8 –10 minutes.
S140 Smarr

Equipment needed. Pencil or pen to record measures. The GDS more consistently differentiates
responses. depressed from non-depressed seniors than other
depression measures (4).
Cost/availability. Available from the original
Yesavage et al (2) article; English long and short Construct. Yesavage and colleagues (2) validated
versions, scoring instruction, and versions in many the 30-item version using two depressive
languages available at www.Stanford.edu/ symptomatology measures, the Zung Self-rating
⬃yesavage/GDS.html. There is no cost, it is in Scale for Depression (SDS) and the Hamilton
public domain. Rating Scale for Depression (HRSD), to compare
their ability to classify normal subjects from mild
Scoring and severe depression. The measures yielded
Responses. Scale. Yes or no. similar results, with normal subjects scoring lower
than persons endorsing mild depressive symptoms
Score range. The range is 0 (no depression) to 30 and those endorsing severe depressive symptoms,
(severe depression) for long form, 0 (no depression) and persons with severe symptoms having the
to 15 (severe depression) for short form. highest scores. When compared to a diagnostic
classification variable, the GDS and HRSD yielded
Interpretation of scores. Higher GDS scores are similar results, while the SDS appeared to
indicative of more severe depression. Brink et al discriminate less effectively. Correlation between
(1) suggested GDS scores 1–10 be considered the GDS and the SDS was 0.84; correlation
normal, while GDS ⱖ11 indicative of possible between the GDS and the HRSD was 0.83.
depression. Using a cut-off score of 14 avoids false Other studies have used depression measures
positives. (i.e., CES-D) to examine the GDS convergent
validity. Stiles and McGarrahan (4) reported that
Method of scoring. Total score calculated by most studies report correlations ranging from 0.58
summing responses that endorse depression; to 0.89. Studies involving young subjects reported
Negatively endorsing items 1, 5, 7, 9, 15, 19, 21, lower correlations (0.66 – 0.67).
27, and 29 indicate depression, while positively
endorsing the remaining 20 items indicates Divergent. The correlations between the GDS
depression. and cognitive screening tests and Mini Mental
State Examination and modified Blessed Test were
Time to score. Two minutes. low since intended to measure different constructs.

Training to score. Minimal, 5 minutes. Sensitivity/responsiveness to change. Sensitivity

of the GDS to change was compared to the SDS
Training to interpret. Minimal, 5 minutes. and HRSD; normal subjects were expected to
receive the lowest GDS scores, and persons
Norms available. None. reporting with severe depressive symptoms to
receive the highest scores. When SDS, HRSD, and
Psychometric Information GDS scores were compared, the 3 severity levels
Reliability. Internal consistency. High: seen in GDS were also seen in criterion measures.
Cronbach’s alpha was 0.94 and split-half reliability
was 0.94.
Comments and Critique
Test-retest reliability. Correlation (r ⫽ 0.85) at 1- The GDS has a simple format that accurately
week retest suggested the GDS scores reflect stable and efficiently assesses depressive symptoms in
individual differences. the elderly, from the young old 65–74 to the oldest
old age 85⫹. A gap remains regarding the validity
Validity. Content. Items are based on of the GDS in persons age 85⫹. The GDS appears
characteristics of depression in the elderly. Brink valid in younger samples, yet may not be the best
and colleagues (1) selected 100 items that choice of assessment with a younger sample. The
distinguished between elderly depressed and non- GDS may also assess “general distress” rather than
depressed individuals; 30 items were selected for only depressive symptoms — several items are
GDS using an empirical selection procedure. indicative of both cognitive and somatic symptoms
of anxiety.
Criterion. High correlations have been noted GDS Short Form and Long Form were highly
between GDS and other depressive symptom correlated (0.84) with and sensitive to depressive
Depression S141

symptoms in mild to moderate dementia. Debate in Depression Inventory as screening instruments in

the literature concluded that the GDS was effective an older adult outpatient population. Psychol
and reliable with individuals with mild dementia. Assess 1992;4:190 –92.
Stiles and McGarrahan (4) recommend caution Rule BG, Harvey HZ, Dobbs AR. Reliability of the
when using the GDS with cognitively impaired Geriatric Depression Scale for younger adults. Clin
individuals, and also recommend not using the Gerontol 1989;9:37– 43.
scale with severely cognitively impaired patients Sheikh JI, Yesavage JA, Brooks JO, III, Freidman L,
Gratzinger R, Hill RD, et al. Proposed factor
or individuals with compromised insight and
structure of the Geriatric Depression Scale. Int
accuracy of self-report.
Psychoger 1991;3:23– 8.
Simple administration and robust
psychometric properties have led to the GDS being
translated into many languages and cultures, yet
studies conducted in other countries/cultures HOSPITAL ANXIETY AND DEPRESSION
suggest that depressive symptoms are expressed SCALE (HADS)
differently in other parts of the world, suggesting
cautious use. General Description
Stiles and McGarrahan (4) recommend the Purpose. To assess anxiety and depressive
following: 1) use oral administration since it is symptoms in a general medical population (1, 2).
appropriate for persons with the widest range of
abilities, 2) use long form versus short form since Content. There are 7 depression items measuring
it is more reliable and valid, 3) use cutoff scores of cognitive and emotional aspects of depression,
11 for higher sensitivity and 14 for higher predominately anhedonia, intermingled with 7
specificity for screening efficiency, and 4) use anxiety items that focus on cognitive and
cautiously with cognitively impaired populations emotional aspects of anxiety. Somatic items
and with non-white populations (more data relating to emotional and physical disorders are
needed). excluded.

References Developer/contact information. A. S. Zigmond

1. (Original) Brink TL, Yesavage JA, Lum O, Heersema and R. P. Snaith, St James’ University Hospital at
PH, Adey M, Rose TL. Screening tests for geriatric Leeds Leeds, UK.
depression. Clin Gerontol 1982;1:37– 43.
2. Yesavage JA, Brink TL, Rose TL, Lum D, Huang V, Versions. Available in English, as well as all
Adey M, et al. Development and validation of a other languages of Western Europe and many of
geriatric depression screening scale: a preliminary Eastern Europe, Scandinavia along with some
report. J Psychiatr Res 1983;17:37– 49. African and Far East languages, including Arabic,
3. Sheikh JI, Yesavage JA. Geriatric Depression Scale
(GDS): recent evidence and development of a shorter
Chinese, Danish, Dutch, Finish, French, German,
version. Clin Gerontol 1986;5:165–73. Hebrew, Italian, Japanese, Norwegian, Portuguese,
4. Stiles PG, McGarrahan JF. The Geriatric Depression Spanish, Swedish, Thai, and Urdu.
Scale: a comprehensive review. J Clin Geropsychol
1998;4:89 –110. Number of items in scale. There are 14 items.

Subscales. Anxiety subscale (HADS-A) and

Additional References
Depression subscale (HADS-D).
Alden D, Austin C, Sturgeon R. A correlation between
the Geriatric Depression Scale Long and Short
Forms. J Gerontol 1989;44:124 –5.
Populations. Developmental/target. General
Harper RG, Kotik-Harper D, Kirby H. Psychometric medical outpatients, ages 16 – 65.
assessment of depression in an elderly generally
medical population. J Nerv Ment Dis 1990;178:113– Other uses. Used extensively, primarily with
9. psychiatric and medical patients, as well as the
Katz PP, Yelin EH. Prevalence and correlates of general population, students, non-patients, and
depressive symptoms among persons with subjects with chronic medical conditions.
rheumatoid arthritis. J Rheumatol 1993;20:790-6.
Herrmann (3) tabulated HADS literature specifying
Montorio I, Izal M. The Geriatric Depression Scale: a
review of its development and utility. Int
study type, medical specialty, population, and
Psychogeriatr 1996;8:103–12. originating country where validated.
Olin JT, Schneider LS, Eaton EM, Zemansky MF, Pollock
VE. The Geriatric Depression Scale and the Beck WHO ICF components. Participation restriction.
S142 Smarr

Administration HADS-A and from 0.82 to 0.90 for HADS-D.

Method. Paper and pencil self-administered Cronbach’s alpha for the German version was 0.80
questionnaire. for HADS-A and 0.81 for HADS-D. Similar or
slightly lower coefficient alphas were observed for
Training. None, designed as easy, short, and to other versions.
be administered in clinic.
Test-retest. High test-retest correlations (r ⬎
Time to administer/complete. 5–10 minutes. 0.80) were found after ⱕ2 weeks and gradually
decrease as time lapses (2– 6 weeks 0.73– 0.76, and
Equipment needed. Pencil or pen to endorse ⬎6 week 0.70).
items.
Validity. Content. The HADS relies on
Cost/availability. Copyrighted and available anhedonia, not on somatic complaints, and is
from: Nfer-Nelson, Darville House, 2 Oxford Road sensitive to mild distress. Construction of the
East, Windsor, Berkshire, SL4 IDF, UK. A test HADS depression subscale minimizes effect of
manual (2) accompanies the scale and describes somatic disorders associated with depression.
administration, scoring procedures, and
psychometrics. Available from the following Web Concurrent. Correlations with corresponding
sites: www.clinical-supervision.com/hads/htm measures of the same theoretical construct (i.e.,
www.organon.se/deprimerad/allm/had/hadskala.htm, anxiety or depression) were adequate. Significantly
www.psychiatry.ox.ac.uk/cebmh/guidelines/ higher correlations were found between HADS-D
depression/appendix1.html#had, and observer ratings and self-assessments for
www.sahs.utmb.edu/Psychology/ depression than with observer and self-ratings of
Adultrehab/hospital_anxiety_and_depression_.htm. anxiety; similar finding identified with measures of
You can request a manual from author at no cost. anxiety and the HADS-A. Compared to commonly
used depression and anxiety measures (BDI, State-
Scoring Trait Anxiety Inventory, Symptom Checklist-90-
Responses. Scale. The scale is a 4-point Likert Revised), correlations with the HADS-D and
Scale, ranging from 0 to 3. HADS-A ranged between 0.60 (good) to 0.80 (very
good) (4).
Score range. Range 0 – 42 for the total score; 0 –
21 for HADS-A and HADS-D. Predictive. HADS has been found to relate to
prospective measurements in renal transplant
Interpretation of scores. Higher scores indicate patients, baseline HADS predicted followup
greater severity. Zigmond and Snaith (1) psychiatric morbidity using a clinical interview. In
recommended the following cutoff scores for the coronary patients, baseline HADS-D predicted
subscales: 0 –7 considered non-case, 8 –10 quality of life ⱖ2 years later, and HADS-D
considered possible case, 11–21 considered predicted surgical outcomes in urology patients.
probable case.
Discriminant. The correlation between HADS-A
Method of scoring. Sum the ratings of 14 items and HADS-D average 0.56 (range 0.49 – 0.74).
to yield a total score; sum the rating on 7 items on
each subscale to yield separate scores for anxiety Sensitivity/responsiveness to change. Designed
and depression. to identify probable “cases” of anxiety or
depression. The HADS is not a diagnostic tool; it is
Time to score. 1–2 minutes. a poor predictor of making a specific diagnosis (5).
Average sensitivities and specificities are ⱖ0.80,
Training to score. Minimal. similar to other self-rating screening tools (3,4).
Bjelland et al’s (2001) tabulation estimated HADS
Training to interpret. Minimal. sensitivity and specificity at optimal cutoffs (4).
Silverstone (5) and Goldberg (6) compared HADS
Norms available. None.
depression subscale scores with standard clinical
assessments in medical patients. Sensitivity
Psychometric Information estimates ranged from 56 to 100%, and specificity
Reliability. Internal consistency. Cronbach’s estimates ranged from 73 to 94%. Positive
coefficient alpha ranges from 0.78 to 0.93 for predictive values range from 19 to 70%. These
Depression S143

estimates favorably compares to studies using BDI Snaith RP, Zigmond AS. The Hospital Anxiety and
and CES-D. Depression Scale. BMJ (Clin Res Ed) 1986;292:344.
HADS scores are responsive to pharmacologic
and psychotherapeutic interventions (3).
PRIMARY CARE EVALUATION OF
Comments and Critique MENTAL DISORDERS-MOOD MODULE
HADS is a reliable, valid method for assessing (PRIME-MD)
emotional distress in medical populations. Despite General Description
its brevity, it screens for possible anxiety and Purpose. A 2-stage (screening and structured
depressive symptompatology similar to more interview) diagnostic instrument designed for
comprehensive clinical measures. The HADS can primary care physicians in general medical settings
be used in clinical and research settings, and may to identify persons with mental disorders (1). The
be particularly useful when studying the cognitive Mood Module was developed to guide the
processes associated with depressive symptoms clinician to a criterion-based diagnosis of
and anxiety, since it is free of physical symptoms, depressive disorders based on the American
such as insomnia and weight loss. The HADS has Psychiatric Association’s Diagnostic and Statistical
good psychometric properties, making it a good Manual of Mental Disorders, Revised Third Edition
choice to measure psychological distress, to (DSM-III-R), now updated to DSM-IV.
differentiate symptoms of depression and anxiety,
or to examine the impact of cognition on Content. PRIME-MD consists of 2 components: a
depression or anxiety (4). 27-item patient questionnaire (PQ) to be used as a
screening tool, and a clinician evaluation guide
References (CEG) that serves as a structured interview to
evaluate mental disorders in 5 diagnostic areas.
1. (Original) Zigmond AS, Snaith RP. The Hospital
The Mood Module in the CEG can be
Anxiety and Depression Scale. Acta Psychiatr Scand
1983;67:361–70.
administered independent of the PQ. Mood
2. Snaith RP, Zigmond AS. The Hospital Anxiety and Module items directly reflect the DSM-III-R
Depression Scale Manual. Windsor, Berkshire (UK): diagnostic criteria for a particular mood disorder
Nfer-Nelson; 1994. and assess sleep disturbance, appetite, anhedonia,
3. Hermann C. International experiences with Hospital low self-esteem, depressed mood, concentration
Anxiety and Depression Scale: a review of validation difficulties, suicide ideation, psychomotor agitation
data and clinical results. J Psychosomatic Res 1997; or retardation, and fatigue.
42:17– 41.
4. Bjelland I, Dahl AA, Haut TT, Neckelmann D. The Developer/contact information. Robert L.
validity of the Hospital Anxiety and Depression Scale: Spitzer, MD, of the Biometrics Research
an updated literature review. J Psychosomatic Res
Department, New York State Psychiatric Institute
2002;52:69 –77.
5. Silverstone PH. Poor efficacy of the Hospital and
and Columbia University.
Anxiety Depression Scale in the diagnosis of major
depressive disorder in both medical and psychiatric Versions. Spanish, Danish, and German
patients. J Psychosomatic Res 1994;38:441–50. versions. PRIME-MD has been shortened to
6. Goldberg D. Identifying psychiatric illness among maximize clinical usefulness. The 2 components of
general medical patients. BMJ 1985;29:161–2. the original PRIME-MD were combined into a 3-
and 4-page, self-administered version called
PRIME-MD Patient Health Questionnaire (PHQ) (2).
Additional References A 2-page version, the Brief PHQ, has also been
Johnston M, Pollard B, Hennessey P. Construct developed. PHQ-9 is the depression module of
validation of the Hospital Anxiety and Depression PHQ.
Scale with clinical populations. J Psychosomatic
Res 2000;48:579 – 84. Number of items in scale. The Mood Module
Lisspers J, Nygren A, Soderman E. Hospital Anxiety and
consists of 17 questions pertaining to depressive
Depresion Scale (HAD): some psychometric data for
a Swedish sample. Acta Psychiatr Scand 1997;96:
symptoms. Thirteen items require the interviewer
281– 6. to elicit responses, 4 items require an interviewer
Mykletun A, Stordal E, Dahl AA. Hospital Anxiety and response. 10 items pertaining to past 2 weeks
Depression Scale: factor structure, item analysis assess for major depressive disorder (MDD), 1 item
and internal consistency in a large population. Br J assesses for partial remission or reoccurrence of
Psychol 2001;179:540 – 4. MDD, 2 items address symptoms over past 2 years
S144 Smarr

to assess for dysthymic disorder (DD), 2 items Equipment needed. None.

assess symptoms of minor depressive disorder
(MND), 1 item assesses the need to rule out bipolar Cost/availability. PRIME-MD is a trademark of
disorder (BD), and 1 item assesses the need to rule Pfizer, Inc. All versions are distributed free to
out depressive disorder due to physical disorder, healthcare providers by local Pfizer sales
medication, or other drug. representatives, and may be photocopied ad
libitum. Pfizer, Inc. provides a CD-Rom containing
Subscales. Mood module consists of 6 sections the PHQ and the Brief PHQ.
pertaining to the following diagnostic categories:
MDD, DD, partial remission or reoccurrence of Scoring
MDD, MND, rule out depressive disorder due Responses. Scale. Yes/No.
physical disorder, medication, or other drug, or
rule out bipolar disorder. Score range. No score; diagnostic categories
result.
Populations. Developmental/target. Original
sample consisted of 1,000 adults from 4 outpatient Interpretation of scores. Three diagnostic
settings with 8 medical diagnoses, (i.e., arthritis, categories in the PRIME-MD system include DSM-
hypertension, diabetes, heart disease, and IV diagnoses of MDD, DD, and partial remission or
pulmonary disease). PRIME-MD was validated by a reoccurrence of MDD. The Mood Module includes
telephone re-interview of a subsample of 431 subthreshold diagnoses (i.e., MND) since
individuals using the Structured Clinical Interview psychiatric symptoms below DSM-IV diagnostic
for DSM-III-R (SCID). categories are associated with functional
impairments and individuals may benefit from
Other uses. Validity data available for Spanish monitoring or treatment. Two rule out diagnoses
and German versions. Self-administered are included (i.e., rule out depressive disorder due
paper/pencil and computerized versions available. physical disorder, medication, or other drug and
The PRIME-MD has been used in specialty clinics rule out bipolar disorder). More than 1 diagnosis
(rheumatology, neurology, plastic surgery, can be made, such as MDD and DD.
oncology, and endocrinology), with veteran
populations, and in rural settings. The developers Method of scoring. PRIME-MD questions require
provide validity data for obstetrics-gynecology a yes or no response. The interview uses a self-
outpatients. Validity data are also available for explanatory decision tree to guide the interviewer
urban American Indians in primary care, but in a systematic manner.
studies of reservation populations are needed.
Time to score. Following tree structure, the
WHO ICF Components. Participation restriction. PRIME-MD is scored during the interview, which
takes ⬍10 minutes.
Administration Training to score. Minimal, allow time to
Method. Brief, straightforward, and easy to understand the decision tree prior to
administer. The PQ is a paper and pencil screening administering.
tool to determine if administration of the PRIME-
MD mood module is indicated. The Mood Module Training to interpret. Minimal training is
is interviewer administered and can be required for health professionals who can provide
administered without the PQ if clinical appropriate psychotherapeutic intervention and
information indicates depressive disorders referrals to diagnosed individuals. Clinical
assessment. supervision may be needed; interviewers may need
to provide individuals meeting the criteria for
Training. Minimal, although should be depressive disorders with treatment approaches
administered and scored by a professional who has (pharmacologic and/or psychological), including
familiarity with or interest in psychiatric referral options. Suicide risk associated with
diagnostic categories. depression must be taken seriously and promptly
addressed. Plans must exist to immediately deal
Time to administer/complete. Authors report with anyone who is an imminent danger to self or
mean completion time of CEG portion as 8.4 others.
minutes; administration time of PRIME-MD mood
module ⬍10 minutes. Norms available. No.
Depression S145

Psychometric Information researchers an easier method to diagnosing mental

Reliability. Interrater reliability. Agreement disorders and offers a less expensive and time-
between primary care physicians (PCPs) and effective approach to diagnosing depressive
mental health professionals (MHPs) regarding disorders. Individuals seeking medical care ought
reinterviewed patients was used to demonstrate to receive earlier identification and treatment,
interrater reliability (1). Kappa coefficients were thereby reducing disability and unnecessary
used to represent agreement, correcting for suffering.
agreement due to chance. Agreement for all The PRIME-MD has been reduced to
psychiatric diagnoses was good (0.71) and was administering 4 core items (sleep disturbance,
satisfactory for the mood module (0.61). Diagnostic anhedonia, low self-esteem, and decreased
prevalence rates for PCPs and the MHPs did not appetite) following the administration of the 2
identify systematic over- or under-diagnosing. depression screening items on the PQ. This
approach is called the “S4 model” and has begun
Validity. Content. Items developed directly from to emerge using rheumatology patients (3).
DSM-III-R criteria, thereby a diagnostic tool. With the PRIME-MD methodology available in
brief formats, many clinicians/researchers will be
Criterion. A SCID administered by MHPs was able to identify a version that suits their individual
considered to be the “gold standard” measure to needs and settings.
assess the validity of the PCPs’ PRIME-MD
evaluation (1). The sensitivity (proportion of cases References
given a diagnosis by MHPs who were correctly 1. (Original) Spitzer RL, Williams JBW, Kroenke K,
identified by the PCPs) was very good for any Linzer M, deGruy FV III, Hahn SR, et al. Utility of a
psychiatric disorder (83%). The Mood Module was new procedure for diagnosing mental disorders in
found to be satisfactory (67%), ranging from 22% primary care: the PRIME-MD 1000 study. JAMA 1994;
for MND to 57% MDD. The specificity (proportion 272:1749 –56.
of cases identified by MHPs not given the 2. Spitzer RL, Kroenke K, Williams JBW, the Patient
diagnosis correctly identified by the PCPs) was Health Questionnaire Primary Care Study Group:
Validation and utility of a self-report version of
excellent across diagnostic modules (88%) and for
PRIME-MD: the PHQ (Patient Health Questionnaire)
the mood module (92%, ranging from 94% for primary care study. JAMA 1999;282:1737– 44.
MND to 98% for MDD), indicating that the PRIME- 3. Jackson JL, O’Malley PG, Kroenke K. Clinical
MD seldom resulted in diagnoses not confirmed predictors of mental disorders among medical
using the SCID. Overall accuracy rate was excellent outpatients: validation of the “S4” model.
across modules (86%) and mood module (84%). Psychosomatics 1998;39:431– 6.

Sensitivity/responsiveness to change. Using Additional References

SCID as the criterion measure, the PRIME-MD
Brody DS, Hahn SR, Spitzer RL, Kroenke K, Linzer M,
reflect changes identified during a standard clinical
deGruy FV III, et al. Identifying patients with
interview. depression in the primary care setting: a more
efficient method. Arch Intern Med 1998;158:2469 –
Comments and Critique 75.
The PRIME-MD has streamlined the Hahn SR, Kroenke K, Williams JBW, Spitzer RL.
Evaluation of mental disorders with the PRIME-
cumbersome psychiatric nomenclature for
MD. In: Maruish ME, editor. Handbook of
untrained health professionals. Even though based psychological assessment in primary care settings,
on DSM-III-R criteria, the mood module remains 2nd edition. Mahwah (NJ): Lawrence Erlbaum
current since the updated DSM-IV depressive Associates; 2000. p. 191–253.
disorder criteria changed little from DSM-III-R Kroenke K, Spitzer RL, Williams JBW. The PHQ-9:
criteria. validity of a brief depression severity measure.
The PRIME-MD offers clinicians and J Gen Intern Med 2001;16:606 –-13.
 S146

Summary Table for Depression Measures*

Psychometric properties
Measure/ Number of Response Method of Time for Validated
scale Content Measure outputs items format administration administration populations Reliability Validity Responsiveness

BDI Cognitive, affective, Total score; cognitive- 21 items 0–3 rating Self 5–10 minutes Psychiatric and Excellent Good Adequate
somatic, and affective (13 items) scale normal
vegetative items and somatic- populations
performance (8-
items) subscales
CES-D Positive affect, negative Total score 20 items 4-point Self ⬍10 minutes General Excellent Excellent Good
affect, activity level, Likert population,
and interpersonal scale including RA
items and fibromyalgia
GDS Affective and cognitive Total score 30-long form Yes/No Self 8–10 minutes Elderly, Excellent Good Good
symptoms common 15-short form hospitalized and
in elderly in the
community
HADS Intermingled Depression and anxiety 14 items 4-point Self 5–10 minutes General medical Good Good Good
depression and subscales; 7 items Likert outpatients
anxiety items each scale
PRIME-MD Items correspond to Diagnostic categories of 17 items in Yes/No Interview ⬍10 minutes Eight medical Good Good Good
the DSM-III-R major depressive Mood diagnoses,
diagnostic criteria of disorder (MDD), Module including
six mood disorders dysthymic disorder, arthritis
partial remission or
occurrence of MDD,
minor depressive
disorder, rule-out
depressive disorder
due to physical
disorder, medication,
or other drug, and
rule out bipolar
disorder

* BDI ⫽ Beck Depression Inventory; CES-D ⫽ Center for Epidemiological Studies-Depression Scale; GDS ⫽ Geriatric Depression Scale; HADS ⫽ Hospital Anxiety and Depression Scale; PRIME-MD
⫽ Primary Care Evaluation of Mental Disorders–Mood Module.
Smarr