Format

Send to

Choose Destination
Psychol Assess. 2019 Feb;31(2):248-264. doi: 10.1037/pas0000665. Epub 2018 Nov 1.

Measurement equivalence of PROMIS depression in Spain and the United States.

Author information

1
Health Services Research Unit, Institut Hospital del Mar d'Investigacions Mèdiques.
2
Centre de Salut Mental d'Adults Sant Martí Sud, Institut de Neuropsiquiatria i Addiccions, Parc de Salut Mar.
3
Centre d'Atenció i Seguiment Barceloneta, Institut de Neuropsiquiatria i Addiccions, Parc de Salut Mar.
4
Centre de Salut Mental d'Adults Sant Martí Nord, Institut de Neuropsiquiatria i Addiccions, Parc de Salut Mar.
5
Centre Fòrum, Institut de Neuropsiquiatria i Addiccions, Parc de Salut Mar.
6
Centre de Salut Mental d'Adults Martí i Julià, Institut de Neuropsiquiatria i Addiccions, Parc de Salut Mar.
7
Programa Trastorn Límit de la Personalitat, Institut de Neuropsiquiatria i Addiccions, Parc de Salut Mar.

Abstract

In this study, we assessed the psychometric properties of the Spanish Patient-Reported Outcomes Measurement Information System (PROMIS) Depression in an adult population-representative sample from Spain (n = 1,503). We tested unidimensionality and local independence item response theory (IRT) assumptions with confirmatory factor and bifactor models under the exploratory structural equations modeling framework. We evaluated item monotonicity assumption with Mokken scaling analysis. We calibrated the items with an IRT-graded response model and assessed score reliability and test information, and evidence of validity with regard to scores on external measures. To examine differential item functioning by age, sex, education, and country (United States vs. Spain, N = 2,271), we used ordinal logistic regression. Results support compliance with IRT assumptions. We found few signs of differential item functioning: Only one item showed country differential functioning between the United States (n = 768) and Spain, with minimal impact on the overall score. Information values were equivalent to reliabilities over 0.90 from -1 (low depression) to +4 SD (high depression) around the population score mean. Evidence of validity in relation to concurrent measures was supported by the expected correlation pattern with external variables of depression, but higher than expected correlations with anxiety were found. Results indicate that the Spanish version of PROMIS Depression is adequate for assessing and monitoring depression levels in the general population and that PROMIS Depression is especially suitable for cross-national comparisons. (PsycINFO Database Record (c) 2019 APA, all rights reserved).

PMID:
30382716
DOI:
10.1037/pas0000665
[Indexed for MEDLINE]

Supplemental Content

Full text links

Icon for American Psychological Association
Loading ...
Support Center