A longitudinal evaluation of the Center for Epidemiologic Studies-Depression scale (CES-D) in a rheumatoid arthritis population using Rasch analysis

Abstract

Background: The aim of this study was to test the internal validity of the total Center for Epidemiologic Studies-Depression (CES-D) scale using Rasch analysis in a rheumatoid arthritis (RA) population.

Methods: CES-D was administered to 157 patients with RA over three time points within a 12 month period. Rasch analysis was applied using RUMM2020 software to assess the overall fit of the model, the response scale used, individual item fit, differential item functioning (DIF) and person separation.

Results: Pooled data across three time points was shown to fit the Rasch model with removal of seven items from the original 20-item CES-D scale. It was necessary to rescore the response format from four to three categories in order to improve the scale's fit. Two items demonstrated some DIF for age and gender but were retained within the 13-item CES-D scale. A new cut point for depression score of 9 was found to correspond to the original cut point score of 16 in the full CES-D scale.

Conclusion: This Rasch analysis of the CES-D in a longstanding RA cohort resulted in the construction of a modified 13-item scale with good internal validity. Further validation of the modified scale is recommended particularly in relation to the new cut point for depression.

Figure 1

Person item distribution graph for 13-Item CES-D scale.

Figure 2

Item characteristic curve for item 10 DIF for age groups.

Figure 3

Item characteristic curve for item 17 DIF for age groups.

Figure 4

Item characteristic curve for item 10 DIF for gender.

Figure 5

Item characteristic curve for item 17 DIF for gender.

Figure 6

Boxplot comparing cut point of 9 on CES-D 13 with the original cut point of 16 on CES-D 20. The recommended cut point of 9 is shown as a horizontal line.