Diagnostic accuracy of a Brazilian depression self-report measure (EBADEP): Original and short versions

Abstract

The present study aimed to investigate the diagnostic accuracy of a scale for the assessment of depressive symptoms, the EBADEP-A, and its short version. Also, we propose the application of the concept of equivalence to compare the discriminative capacity between the two scales, i.e., to determine whether a new procedure (i.e., EBADEP-A short version) is not worse than the procedure in use (i.e., the EBADEP-A). Participants were 80 individuals, 40 with a diagnostic of major depressive disorder based on SCID-I, and 40 healthy controls. Age mean was about 41, and most participants were female (85%). We calculated sensitivity, specificity, positive and negative likelihood ratio, positive and negative predictive power, and negative predictive power (PP−), based on the ROC curve for the EBADEP-A full and short versions. Both versions of EBADEP-A proved useful for the diagnostic of major depression. We concluded that the short version of EBADEP-A is closely equivalent to its extended version. Based on the results we found, we conclude that using diagnostic accuracy combined to the concept of equivalence can not only assist professionals with a focus on psychometric parameters, but also have an impact on the application of the tools in clinical settings.

Diagnosis Accuracy of Baptista Depression Scale: Adult and Screening Versions

“…Moreover, the results suggest that the screening version, with a smaller number of items, has a capacity similar to the full version of EBADEP-A. This finding is in line with what was observed by Baptista and Carvalho (2018), who also found equivalence between the versions. The results of the diagnostic accuracy indicators, in general, also suggest the use of the screening version as this version obtained similar results to the full version, using fewer items and requiring less administration time.…”

“…A rating scale is used to measure if the participant agrees more with the first statement, the second, or with both at the same magnitude. Studies have revealed good psychometric properties for the scale (Baptista, 2012;Baptista & Carvalho, 2018;Bighetti et al, 2014).…”

“…Studies carried out using EBADEP-A observed evidence for its use and good psychometric properties (Baptista, 2012;Baptista et al, 2012;Bighetti et al, 2014;Carneiro & Baptista, 2012;Souza et al, 2015). In addition to EBADEP-A, a screening version was developed, with 15 items on a 4-point Likert scale (Baptista & Carvalho, 2018). Shorter scales can provide benefits, such as lower spending and faster assessments, and are useful tools mainly in public health services and primary care (Bolsoni & Zuardi, 2015).…”

“…The first study encompassed 1,676 individuals, observing sensitivity equal to 77.5% and specificity equal to 87.5 (Baptista, 2012); and another study with 40 depressive patients (MDE or MDD) and 40 people from the general population and without a diagnosis, observing sensitivity equal to 97.5% and specificity equal to 90.5% (Baptista & Carvalho, 2018). In this same study (Baptista & Carvalho, 2018), the short version of EBADEP-A (EBADEP-screening) was also used, which obtained 95% sensitivity and 87.5% specificity. Screening tests should have greater sensitivity in comparison to specificity, making it possible to identify all individuals with the condition (Germans et al, 2012).…”

“…Besides, we sought to verify which of the two versions of EBADEP is the most discriminative. In order to achieve the proposed objectives, we replicated the procedures used by Baptista and Carvalho (2018), although in a larger sample. We created four hypotheses: H1) the clinical group should present higher means in EBADEP-A and EBADEP-screening (Baptista, 2012); H2) EBADEP-A and EBADEP-screening should show an area under the curve (ROC curve) > 0.80, considered good (Bewick et al, 2004;Baptista & Carvalho, 2018Hosmer & Lemeshow, 2000; H3) EBADEP-A and EBADEP-screening should have greater sensitivity and less specificity to select a greater number of cases (Germans et al, 2012); H4) EBADEP-A should have greater discriminative power compared to EBADEP-screening (Baptista & Carvalho, 2018).…”