Citation

Thombs BD. Burns 2007; 33(5): 547-553.

Affiliation

Department of Psychiatry and Behavioral Sciences, Johns Hopkins University School of Medicine, Baltimore, MD, USA. brett.thombs@mcgill.ca

Copyright

(Copyright © 2007, Elsevier Publishing)

DOI

10.1016/j.burns.2006.10.398

PMID

17485178

Abstract

The objective of this study was to assess whether scores on the Beck Depression Inventory (BDI) are biased by injury severity among hospitalized survivors of burn (N=262). A confirmatory factor analysis (CFA) model was developed with a general depression factor that loaded on all items and somatic and cognitive factors that were orthogonal to the general factor and to each other. The model fit the data well and substantially better than an alternative three-factor model with correlated factors. Percent total body surface area burned (TBSA) was significantly associated with the general depression factor (p=.04), but also with the orthogonal somatic factor (p<.001), suggesting biased measurement due to overlap between somatic symptoms of depression and the severity of the burn injury. Analysis of item communalities, however, suggested that only approximately 2% of total predicted item variance was associated with bias related to injury severity. It was concluded that, despite a small amount of bias, the BDI is a reasonably accurate clinical tool even in the context of severe burn. Appropriate adjustments for bias, however, should be made in research with the BDI among patients with acute burn.

Language: en