Validation of Three Previously Developed Short Forms of the Beck Depression Inventory for Youth: A Replication

Abstract

Three previously developed short forms of the Beck Depression Inventory-Youth (BDI-Y) were validated against the standard 20-item BDI-Y; 168 adolescent survivors completed the standard and short-form versions of the BDI-Y. The short forms were evaluated for internal consistency and compared with the standard BDI-Y using correlation coefficients and receiver operating characteristic curve analyses. The three short forms had good internal consistency (α > 0.85), high correlations with the total BDI-Y scale (r > 0.85), and good discrimination compared with the standard BDI-Y cutoff score (area under the ROC curve >0.95). Consistent with prior findings, strong psychometric properties of an eight-item short form support its use as a screening measure for adolescent cancer survivors.

Introduction

Treatment for cancer during childhood often presents as a health crisis associated with significant emotional distress and developmental disruption for affected children. Although most childhood cancer survivors do not experience significant adjustment problems after treatment, this population as a whole is at increased risk for psychological symptoms, including depression, and suicide even many years after completion of therapy.^1–3 Because they are vulnerable to adjustment problems, psychological assessment has been widely supported for childhood cancer patients and survivors.^4–7 To be successfully integrated into survivorship care, psychological screening measured for these young survivors will need to be brief and easily administered in a medical setting. Several brief scales (<10 items) for depression have been evaluated in adult survivors,⁸ but similar brief measures have not been widely evaluated for adolescent survivors of childhood cancer.

To address this, we investigated the validity of three previously developed brief versions of the Beck Depression Inventory-Youth (BDI-Y).⁹ In a prior study,¹⁰ we reported that 6-, 8-, and 11-item short-form versions of the BDI-Y were highly correlated with the standard 20-item BDI-Y and should be further studied for evaluating depression in adolescent survivors. However, a limitation of that study involved the method in which the three BDI-Y short-form scores were calculated. Rather than using separate administrations of the BDI-Y and the short forms, the BDI-Y was only administered once as part of standard clinical care and this single administration served as the basis for all scores utilized in the study. In the present study, we set out to replicate findings from our previous report and to address this limitation by using separate administrations of the standard BDI-Y and the short-form items.

Materials and Methods

Participants and procedure

Participants were 168 cancer survivors, 82 males (49.0%) and 86 females (51.0%), between the ages of 12 and 18 years (mean = 14.88, standard deviation [SD] = 1.7), followed at a single cancer center. Time since cancer diagnosis ranged from 3.23 to 17.56 years (mean = 10.70, SD = 3.12). Survivors' age at first diagnosis ranged from birth to 13.48 years (mean = 4.19, SD = 3.24) and their cancer diagnoses included the following: leukemia (30.4%), brain tumor (19%), neuroblastoma (19%), sarcoma (9.5%), Wilms tumor (7.7%), non-Hodgkin lymphoma (7.1%), and other cancers (7.1%). In addition, 80.4% of survivors received chemotherapy, 53.6% received surgery, 26.2% received radiation, and 8.3% received a transplant as cancer-related therapy. Survivors completed the standard BDI-Y as part of a packet of intake forms administered by clinic staff before their scheduled medical appointment. After these completed forms were returned to clinic staff, research staff invited survivors (and parents/guardians) to participate in this study. Participants who consented (or had parental consent) completed a second abbreviated depression checklist measure made up of 12 BDI-Y items; the three BDI-Y short forms (6-, 8-, and 11-item short forms) under study were analytically extracted from these 12 items. This second checklist was administered at the same clinic visit, typically within an hour of the intake forms being completed. During the study period, only eight survivors declined to participate. Procedures were approved by the cancer center's institutional review board.

Beck Depression Inventory for Youth

The BDI-Y⁹ is designed to identify symptoms of depression in children and adolescents, as described in the Diagnostic and Statistical Manual of Mental Health Disorders, Fourth Edition.¹¹ It includes 20 Likert scale items asking about depressed thoughts and feelings, feeling of hopelessness, and physiological indications of depression. The 6-, 8-, and 11-item short forms studied were derived and evaluated in a prior study of 202 adolescent cancer survivors.¹⁰ Item numbers indicating specific BDI-Y items included in each of the three short forms are included in Table 1. All together, these three short forms comprise 12 of the standard BDI-Y items.

Table 1.

Psychometric Properties of the Three BDI-Y Short Forms

	Original 20-item BDI-Y	6-item SF		8-item SF		11-item SF
Item numbers	1–20	1, 4, 8, 10, 13, 18		1, 4, 8, 10, 13, 18, 15, 20		1, 3, 4, 8, 10, 11, 12, 13, 16, 18, 20
Cronbach's α	0.93	0.86		0.89		0.91
Item-scale correlations	0.41–0.72	0.41–0.75		0.50–0.77		0.49–0.77
Correlation with BDI-Y	—	0.86		0.87		0.90
AUC with BDI-Y	—	0.97		0.99		0.99
Selected cutoff scores		Sensitivity	Specificity	Sensitivity	Specificity	Sensitivity	Specificity
≥4		0.93	0.89	1	0.89	1	0.79
≥5		0.80	0.94	0.93	0.93	1	0.84
≥6		0.60	0.99	0.87	0.95	1	0.88
≥7		0.47	1	0.73	0.99	1	0.92
≥8		0.33	1	0.47	0.99	0.93	0.95
≥9		0.33	1	0.47	1	0.73	0.98
≥10		0	1	0.40	1	0.67	0.99

Scores in bold reflect suggested cutoff levels.

AUC, area under the receiver operating characteristic curve; SF, short form.

Statistical analysis

Short forms were evaluated by calculating internal consistency (Cronbach's α), corrected item-scale correlations, and correlations with the standard BDI-Y measure. Based on the published BDI-Y criteria,⁹ participants were classified as having elevated depression if they had t-scores ≥55 on the standard BDI-Y. This classification of elevated depression versus no elevation in depression on the standard BDI-Y was used as the criterion against which the short forms were evaluated in receiver operating characteristic (ROC) analyses. Both area under the ROC curve (AUC) and sensitivity and specificity for selected short-form cutoff scores were reported. We modified criteria from our prior analysis and set a priori criteria for short forms to meet to be regarded as acceptable¹⁰: these were internal consistency (Cronbach's α) ≥0.80, correlation with the standard BDI-Y (Pearson r) ≥0.85, AUC ≥0.85, and having at least one cutoff score with a sensitivity of ≥0.90 and specificity of ≥0.75.

Results

Each of the three short forms met study criteria for internal consistency, correlation with the standard BDI-Y, and discrimination between survivors with and without elevated BDI-Y depression scores. Similar to our previous analysis,¹⁰ Cronbach's α for the three short forms ranged from 0.86 to 0.91 and their correlations with the standard BDI-Y ranged from 0.86 to 0.90 (Table 1). In ROC analyses, AUC ranged from 0.97 to 0.99, indicating very good discrimination. Each short form had at least one cutoff score that met study criteria for both sensitivity and specificity. For the six-item short form, a cutoff score of ≥4 showed excellent sensitivity (≥0.93) and specificity (≥0.89); for the eight-item short form, cutoff scores of ≥4 and ≥5 demonstrated high sensitivity (≥0.93) and specificity (≥0.89). For the 11-item short form, sensitivity and specificity were acceptable for all cutoff scores between 4 and 8.

Discussion

This study replicates and extends previous research supporting short forms of the BDI-Y for evaluating depression in adolescent cancer survivors. Results were highly consistent with our previous report; the three short forms demonstrated high internal consistency and correlations with the standard BDI-Y score, as well as cutoff scores that were sensitive and specific for detecting significant depression.

Based on all the available evidence, we continue to recommend the eight-item short form with a cutoff score ≥5 (sensitivity 0.93, specificity 0.93) as a potential screen for depression in this population. Having just eight items, this option is considerably shorter than the standard BDI-Y, and with strong screening accuracy, this short form is quite promising for screening adolescent cancer survivors. The six-item short form is the most parsimonious of the three short forms and is also highly correlated with the original BDI-Y. Cutoff scores on the six-item scale have been somewhat less consistent across all analyses to date, leading us to be somewhat less confident about recommending it as a screening measure. However, it is likely to be very useful in research applications where it could be valuable as a brief measure to describe group differences in depression rather than to make individual screening decisions. The 11-item short form may prove an attractive option in situations where brevity is less of an issue, although it offers less of a reduction in time and patient burden than the alternative short forms.

In contrast to our prior study, this study used separate administrations of the standard BDI-Y and the short forms, allowing participants' item responses to vary independently over the different forms. This approach is less likely to overestimate the agreement of the short forms with the standard BDI-Y, so the fact that results were highly consistent with the previous report is evidence that those earlier results were not an artifact of study method. Moreover, the fact that all short-form cutoff scores supported in this study were also supported in the prior study is reassuring that they are robust across samples and methods.

It should be noted that these results only reflect the association between the standard BDI-Y and the BDI-Y short forms. Future studies using psychiatric interviews as the criterion measure of depression will be necessary to more comprehensively validate the BDI-Y short forms. In addition, this was a single institution study, making it difficult to discern the extent to which findings are generalizable to a larger and more diverse sample. Finally, while this study focuses on screening for depression, screening for other forms of psychological distress is also warranted. However, despite these limitations, the results have significant implications for evaluating depression in young cancer survivors. Clinical care guidelines for pediatric cancer survivors⁵ emphasize the importance of screening for psychological late effects, including depression, but this cannot be achieved without practical methods that minimize patient burden and still provide accurate results. The brief versions of the BDI-Y reported in this study provide practitioners with new options for efficient assessment of depression that can help identify affected adolescent cancer survivors and refer them for needed treatment.

Footnotes

Author Disclosure Statement

No competing financial interests exist.

References

Zeltzer

, Recklitis

, Buchbinder

, et al. Psychological status in childhood cancer survivors: a report from the Childhood Cancer Survivor Study. J Clin Oncol. 2009; 27:2396–404.

Recklitis

, Diller

, Li

, et al. Suicide ideation in adult survivors of childhood cancer: a report from the Childhood Cancer Survivor Study. J Clin Oncol. 2010; 28:655–61.

Schultz

, Ness

, Whitton

, et al. Behavioral and social outcomes in adolescent survivors of childhood cancer: a report from the childhood cancer survivor study. J Clin Oncol. 2007; 25:3649–56.

Recklitis

, O'Leary

, Diller

. Utility of routine psychological screening in the childhood cancer survivor clinic. J Clin Oncol. 2003; 21:787–92.

Children's Oncology Group: Long-term follow-up guidelines for survivors of childhood, adolescent, and young adult cancer. Arcadia, CA: Children's Oncology Group; 2013.

Kazak

, Barakat

, Hwang

, et al. Association of psychosocial risk screening in pediatric cancer with psychosocial services provided. Psychooncology. 2011; 20:715–23.

McCarthy

, McNeil

, Drew

, et al. Psychological distress and posttraumatic stress symptoms in adolescents and young adults with cancer and their parents. J Adolesc Young Adult Oncol. 2016; [Epub ahead of print]; DOI: 10.1089/jayao.2016.0015.

Vodermaier

, Linden

, Siu

. Screening for emotional distress in cancer patients: a systematic review of assessment instruments. J Natl Cancer Inst. 2009; 101:1464–88.

Beck

, Beck

, Jolly

, et al. Manual for Beck Youth Inventories-Second edition for children and adolescents. Oxford: Harcourt Assessments; 2005.

10.

Blackmon

, Liptak

, Recklitis

. Development and preliminary validation of a short form of the Beck Depression Inventory for Youth (BDI-Y) in a sample of adolescent cancer survivors. J Cancer Surviv. 2014; 1–8.

11.

American Psychiatric Association: Diagnostic and statistical manual of mental disorders, 4th Edition, text revision (DSM-IV-TR) (ed 4th). Washington, DC: American Psychiatric Association; 2000.