Beck Depression Inventory-Ii (Bdi-Ii) Is a 21-Item

¶ … Beck Depression Inventory-II (BDI-II) is a 21-item clinician administered and scored scale that is designed to measure a person's mood and symptoms related to depression. The BDI-II was designed to conform to the DSM-IV depression diagnostic criteria and represents a substantial improvement over its predecessor, the original Beck Depression Inventory. The BDI-II has been used both as a research measure (its primary intended use) and to assist with the clinical diagnosis of depression.

The BDI-II has been subject to numerous empirical studies designed to measure its internal consistency, convergent and discriminant validity, criterion validity, and construct validity and the test demonstrates acceptable psychometric qualities, but there have been some concerns with its use. This paper reviews the development of the BDI-II, its psychometric properties, uses, strengths, and weaknesses. Advantages and disadvantages of using the BDI-II and recommendations for future research regarding its use are also discussed.

Title of paper The psychiatric diagnosis of major depressive disorder characteristically begins with the identification of symptoms of which the presence and/or severity that occur over a specific span of time are evaluated. Standardized measures of depression that can be used to measure and document mood, somatic, vegetative, and other depressive symptoms can be useful in the diagnosis of depression and also to determine the severity of the symptoms in the person.

The Beck Depression Inventory (BDI) is a 21-item clinician administered and scored scale designed to measure mood and symptoms related to depression that although developed originally for research purposes in 1961, also enjoyed widespread clinical use (Arbisi, 2001). Following nearly 35 years of clinical and research use the BDI underwent a major revision. The revised version of the Beck, the BDI-II was developed in 1996. This paper reviews the development of the BDI-II, its psychometric properties, uses, strengths, and weaknesses.

Basic Description and Test Development Compared to the original BDI, the BDI-II added items covering such aspects of depression as agitation, worthlessness, concentration difficulties, and loss of energy. Consequently, items were dropped or revised regarding the domains of weight loss, body image change, somatic preoccupation, and work difficulty while still retaining the BDI's 21-item format. The revision was substantial as all but three of the original items were change (Arbisi, 2001).

(Grothe et al., 2005) reports that the revisions of the BDI to the BDI-II were undertaken to make the test correspond more closely to the diagnostic criteria for mood disorders in the DSM-IV by, designing it to correspond to the items of the SCID (Structured Clinical Interview for DSM Disorders) for the DSM-III-R.

Furthermore some items on the original BDI had some issues with clinically failing to differentiate across the range of depression (e.g., mild, moderate, and severe presentations of depression) and several other items were found to display a gender bias (Arbisi, 2001).

In fact, a revision of the original BDI had been developed in 1987 (the BDI-IA) that reworded 15 of the original items 21 items but yet this version still did not address some of the aforementioned issues with the original BDI such as its limited range of depressive symptoms nor its failure to be consistent with DSM diagnostic standards and criteria for mood disorders (Arbisi, 2001). The BDI-II consists of 21 items read by the subject (or alternatively they can be read to the subject by the administrator).

Each item is followed by four options (statements) that the respondent is required to endorse as they are related to their feelings over the prior two weeks including the day of the assessment. The options are scored zero to three, with higher scores reflecting more severe levels of depressive symptomatology. The items reflect different dimensions of depression ranging from sadness to loss of energy to loss of interest in activities such as sex. A test question example is provided below (Beck, Steer, & Brown, 1996): 1. Sadness 0. I do not feel sad. 1.

I feel sad much of the time. 2. I am sad all the time. 3. I am so sad or unhappy that I can't stand it. The time to administer the test typically ranges from five to ten minutes (Arbisi, 2001). The test is designed for and can be administered to individuals 13-86 years of age, provided they are not illiterate (in which case the test can be read aloud to the subject). However it has been used for younger and older subjects as well (Arbisi, 1996).

Following completion of the test the administrator sums up the individual item totals and compares them to standardized cut scores to determine the severity of depression in the individual. The cut scores for the BDI-II were originally developed by classifying 127 University of Pennsylvania outpatients into four groups: mildly depressed (2) moderately depressed (3) severely depressed and (4) nondepressed based on the SCID for the DSM-III. Cut scores were derived through the use of receiver operator characteristic curves (Beck et al., 1996).

The manual for the BDI-II does not provide demographic information for the standardization sample, a potential shortcoming. The BDI-II has been translated into numerous languages and there are computerized and internet versions available (Arbisi, 2001). Psychometric Properties Reliability The BDI-II has demonstrated sound reliability across many different empirically-based studies. According to the BDI-II manual item-total correlations ranged from .38 to .74 (Beck et al., 1996).

According to Arbisi (2001) the internal consistency (Cronbach's alpha) was .92 and .93 respectively for a clinical sample of 500 outpatient therapy patients (91% Caucasian) and a sample of 120 Canadian college students (described as "predominantly" Caucasian). Test-retest over a one-week period was assessed in a very small subsample of 26 of the outpatients and was shown to be high (r = .93).

Over short intervals test-retest reliabilities have been adequate to high ranging from the .70's to .93; however, over longer periods we would not expect this to occur due to the waxing and waning nature of depression (Beck et al., 1996). Hollandare, Andersson, and Engstrom (2010) recruited 87 patients from primary care and psychiatric care in a Swedish public health care system. The participants completed the BDI-II and the Montgomery-Asberg Depression Rating Scale-Self-rated (MADRS-S) on paper versions and versions on the Internet. The order of administration was randomized in order to control for order effects.

The depressive symptom severity in the sample ranged from mild to severe. Cronbach's alpha ranged from good to excellent for the BDI-II regardless of the order of administration (? =.91 for paper administration in the paper-first group; = .87 for internet first group on the internet version and; =.89 for both paper first group completing on the internet version and the internet first group on the paper version). Thus, it appears that the BDI-II has stable internal consistency whether it is administered on paper or on the computer.

Segal, Coolidge, Cahill, and O'Riley (2008) studied the psychometric properties of the BDI-II using a sample of 376 community-dwelling adults with an age range between 17-90 years old. For the entire sample the BDI-II had excellent internal reliability as measured by Cronbach's alpha (? = .90). Alpha was also calculated for young and older adult groups. For the young adult group (17-29 years old) the internal reliability was found to be excellent (? = .92) and for the older adult group (55-90 years old) the internal reliability of was also good (? = .86).

The participants in the aforementioned studies regarding the reliability of the BDI-II were predominately Caucasians. There have been other studies using other ethnic groups as participants that have investigated the reliability of the BDI-II. For example, Grothe et al. (2005) investigated the psychometric properties of the BDI-II using 200 African-American participants with mean age of 49.26 years old (range of 20 to 90 years old). Internal consistency as measured by Cronbach's alpha was consistent with previous findings in Caucasian samples (BDI -- II total score, ? =.90).

These researchers also performed a confirmatory factor analysis on the BDI-II to test the fit of a two-factor model of the BDI-II proposed by other researchers (e.g., Beck et al., 1996). The internal consistency for both factors was also good (Cognitive factor ? = .81; Somatic factor, ? =.87). VanVoorhis and Blumentritt (2008) investigated the psychometric properties of the BDI-II in a sample of 131 Mexican-American youths recruited from three facilities in the Southern Texas border area. Ages of the participants ranged from 13 to19 years, with a mean age of 15.5 years.

Using coefficient alpha the internal consistency of the BDI-II in this sample was consistent with levels found in the aforementioned studies (? = .90). Validity Arbisi (2001) indicated that the BDI-II has good convergent validity reporting that the strong correlations between the BDI-II and BDI-IA (.93), the Beck Hopelessness Scale (.68), the Revised Hamilton Psychiatric Rating Scale for Depression (.71), and the Symptom Checklist-90-Revised (SCL-90-R) Depression subscale (r = .89).

With respect to discriminant validity Arbisi (2001) reported that the correlation between the BDI-II and the Revised Hamilton Anxiety Rating Scale was .47, which is was significantly higher than the correlation between the BDI-II and HRSD-R. However, the BDI-II was moderately correlated with the Beck Anxiety Inventory (r = .60) as well as with the SCL-90-R Anxiety subscale (r = .71), although this last correlation is less than the correlation between the BDI-II and the SCR-90-R depression scale.

In effect substantial correlations between measures of depression and anxiety are consistent with past research investigating the relationship self-reported anxiety and depression (Arbisi, 2001). So perhaps comparing relationships between these concepts is not the best way to determine discriminant validity and indeed such comparisons have been made to establish convergent validity (see below). Segal et al.

(2008) calculated Pearson product-moment correlations between the BDI-II and the Center for Epidemiologic Studies Depression Scale (CES-D); the Coolidge Axis II Inventory (CATI) depression subscale; the CATI anxiety subscale; CATI Depressive Personality Disorder subscale; the Short Psychological Well-Being Scale (SPWB) total score and the SPWB subscales and; the Perceived Stress Scale (PSS), a self-rating scale of overall physical health (scored on a BDI-II demonstrated robust positive correlations with the CATI depression subscale (.58), the CATI anxiety subscale (.53), CATI Depressive Personality scale (.57), CES-D (.68), and the PSS (.67).

Moreover, there was also strong evidence for discriminant validity based on robust negative correlations with the SPWB, an 84-item self-report inventory that measures well-being across six different dimensions. The correlation between the BDI-II and the total SPWB scale was -.65. Correlations with the SPWB subscales were as follows: Autonomy, r = -.33; Environmental Mastery, r = -.62; Personal Growth, r = -.40; Positive Relations with Others, r = -.50; Purpose in Life, r = -.60 and; Self Acceptance, r = -.66.

VanVoorhis and Blumentritt (2007) correlated BDI-II scores with various scales of the Adolescent Psychopathology Scale (APS) completed by their young Hispanic sample. The researchers hypothesized that the BDI-II would be related to subscales measuring depression and internalizing behaviors (convergent validity) and demonstrate weaker relationships with scales measuring externalizing behaviors (discriminant validity). Higher correlations were found with scales that measured depressive disorders or constructs related to depression such as the scale for generalized anxiety disorder (r = .64), major depression (r = .61), dysthymic disorder (r = .62), social phobia (r = .54), and the internalizing disorder factor score (r = .60).

On the other hand, lower correlations were found with scales measuring externalizing behaviors. These scales included conduct disorder (r = .17), oppositional defiant disorder (.30), substance abuse (.18), attention deficit hyperactivity disorder (.41), and the externalizing disorder factor score (.34), generally supporting the researchers' hypothesis. With respect to criterion validity Grothe et al. (2005) compared the scores of the BDI-II obtained by their participants to the results of the Primary Care Evaluation of Mental Disorders (PRIME-MD) Mood Module.

The PRIME-MD Mood Module is a structured interview consisting of yes/no questions used to satisfy the diagnostic criteria for major depressive disorder outlined in the DSM -- IV-TR. They established the presence or absence of major depression in their sample with the PRIME-MD. Further analyses indicated that participants with a diagnosis of major depression had significantly greater BDI -- II total scores compared to participants without the diagnosis (t = 12.83, p < .01).

Uses of the BDI-II The BDI-II was primarily designed for research purposes (Arbisi, 2001; Beck, et al., 1996) but has found its way into clinical use as well. For example regarding research uses, Hollandare, Andersson, and Engstrom (2010) used the BDI-II to help determine if computerized versions of screening instruments for depression retained their psychometric properties compared to the original pencil and paper versions. Their findings indicated that the measures under study were similar whether administered over the internet or paper and pencil.

Other researchers have used the BDI-II to better understand the underlying dimensions of depression via factor analytic models and to understand how depressive symptoms may be differently expressed/experienced by different ethnic groups, ages, or across gender (Grothe et al., 2005; Segal et al., 2008; VanVoorhis and Blumentritt, 2007). With respect to its clinical use Arbisi (2001), Grothe et al. (2005), and Segal et al. (2008) all report that the BDI-II has been one of the most popularly used depression screens in clinical practice.

Hollandare, Andersson, and Engstrom (2010) reported that the BDI-II used as a clinical screen for depression is second only to the Hamilton Rating Scale for Depression. Advantages and Disadvantages of the BDI-II The BDI-II is widely used and accepted as a valid measure of depressive symptomatology (Arbisi, 2001; Grothe et al., 2005). It can be administered orally by an examiner to those with vision problems, reading difficulties or other problems. It is also short, user-friendly, and it is easy to administer and easy to score (Arbisi, 2001).

The BDI-II has been demonstrated to have good internal consistency (Arbisi, 2001; Grothe et al., 2005; Hollandare, Andersson, & Engstrom, 2010; Segal et al., 2008; VanVoorhis & Blumentritt, 2007), acceptable convergent validity (Arbisi, 2001; Grothe et al., 2005; Segal et al., 2008; VanVoorhis & Blumentritt, 2007), good discriminant validity (Arbisi, 2001; Grothe et al., 2005; Segal et al., 2008; VanVoorhis & Blumentritt, 2007), and acceptable criterion validity (VanVoorhis & Blumentritt, 2007). The BDI-II has been translated into languages other than English, and its psychometric properties have been established in numerous cultural groups (Grothe et al., 2005; Segal et al., 2008; VanVoorhis & Blumentritt, 2007).

Moreover, the BDI-II is designed to assess state-related depression and could be used as a quick weekly screener prior to therapy sessions (Grothe et al., 2005). The test has been found to be useful in detecting change in treatment-outcome studies (Segal et al., 2008; VanVoorhis & Blumentritt, 2007). However, clinically due to the face validity of the BDI-II (see test question example above and also Arbisi, 2001; Grothe et al., 2005) its use can result in underreporting or over-reporting of depressive symptoms could this be a concern in clinical use and empirical research.

Moreover, individuals with low education and some Spanish speakers have been shown to experience difficulty with the response format even if the test is read to them (Grothe et al., 2005; VanVoorhis & Blumentritt, 2007). The wording in some items asks the person to compare their current state to a state prior to their current state (e.g., use of terms "than usual" or "as ever").

Individuals with chronic trauma occurring in childhood or chronic low level depression (dysthymia) may respond by circling a zero as they do not feel "worse" than their "usual" state (Hollandare, Andersson, & Engstrom, 2010). Psychometrically there are still several concerns with the BDI-II. The normative sample was predominantly Caucasian (91%) and although the measure is used for adolescents, the norms were collected with mostly adult participants (Beck et al., 1996).

The procedure used to determine the cut scores in the standardization process may increase the likelihood of obtaining false positive diagnoses or over diagnosing of depression clinically and in empirical studies (Grothe et al., 2005). Understanding the construct validity of the BDI-II has proved to be an elusive quarry. Factor analytic studies have yield one, two, and three or more factor solutions and even these results differ over different age groups and ethnicities (Grothe et al., 2005; Segal et al., 2008; VanVoorhis & Blumentritt, 2007).

More research on the construct validity of the BDI-II is needed as well as how the norms apply to different ethnic groups, age groups, and levels of education. Moreover, the majority of psychometric studies conducted with adolescent participants have involved the use of predominantly Caucasian participants and have not included individuals of lower socio-economic status (VanVoorhis & Blumentritt, 2007). Further research on the BDI-II is needed with diverse groups of participants. References Arbisi, P.A. (2001). Review of the Beck Depression Inventory- II. In the fourteen mental measurements yearbook.

Retrieved January 31, 2012, from EBSCO Mental Measurements Yearbook database. Beck, A.T., Brown, G., & Steer, R.A. (1996). Beck Depression Inventory II manual. San Antonio, TX: The Psychological Corporation. Grothe, K.B., Dutton, G.R., Jones, G.N., Bodenlos, J., Ancona, M., & Brantley, P.J. (2005). Validation of the Beck Depression Inventory-II in a low-income African-American sample of medical outpatients. Psychological Assessment, 17, 110-114. Hollandare, F., Andersson, G., & Engstrom, I. (2010).

A comparison of psychometric properties between internet and paper versions of two depression instruments (BDI-II and MADRS- S) administered to clinic patients. Journal of Medical Internet Research, 12(5), e49. Retrieved February, 29, 2012 from http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3057311/. Segal, D.L., Coolidge, F.L., Cahill, B.S., & O'Riley, A.A. (2008). Psychometric properties of the Beck Depression Inventory-II (BDI-II) among community dwelling older adults. Behavior Modification, 32(1), 3 -- 20. VanVoorhis, W.C.R. & Blumentritt, T.L. (2007).

Psychometric properties of the Beck Depression inventory-II in a clinically-identified sample of Mexican-American Adolescents. Journal of Child Family Studies, 16,789-798. Annotated Bibliography Beck, A.T., Brown, G., & Steer, R.A. (1996). This is the test manual for the BDI-II. The authors suggest fully reading the manual before administering the BDI-II. The manual discusses the theory behind depression assessment and the history of the BDI, BDI-IA, and BDI-II. In conjunction with the above the manual also describes the shortcomings of the previous versions of the BDI-II.

The authors make a case for how and why depression assessment should conform to current diagnostic guidelines. Describes the changes in the BDI-II from the BDI that include designing the questions to conform to the SCID. Explains the standardization process used for the development of the test and the procedures used including the ROC curve method. Gives some demographic information for the standardization sample. Gives cut scores for the BDI-II and the instructions on how to administer the test.

Grothe, K.B., Dutton, G.R., Jones, G.N., Bodenlos, J., Ancona, M., & Brantley, P.J. (2005). The researchers acknowledge that the BDI-II has demonstrated good psychometric properties in previous research but point out that one drawback to previous research on the BDI-II is the inclusion of mostly Caucasian samples. The researchers' purpose for conducting this study was to add to the knowledge of the psychometric properties of the BDI-II in African-American participants from lower SES backgrounds.

Participants were African-American participants with mean age of 49.26 years old recruited from medical outpatient clinics who were screened for adequate reading skills with the Woodcock -- Johnson III Oral Comprehension test and then took the BDI-II. Internal consistency of the BDI-II as measured by Cronbach's alpha was consistent with previous findings in Caucasian samples. Six post doctoral students in clinical psychology administered the PRIME-MD Mood Module to the sample. Criterion-related validity for the BDI-II was evaluated with the PRIME-MD and was judged to be adequate for the sample.

To assess construct validity the researchers used confirmatory factor analysis (CFA) in an attempt to replicate the hierarchical two-factor structure identified by Beck et al., 1996 and several other previous studies. The two factors (somatic and cognitive) correlated with each other (r. =.75). The standardized path coefficients of the symptoms for the two factors were salient. The normal theory maximum-likelihood chi-square test for the CFA model was also significant indicating that.