Psychometric properties of the eight-item modified Medical Outcomes Study Social Support Survey based on Spanish outpatients

Abstract

The aim of this study was to examine the psychometric properties of the modified Medical Outcomes Study Social Support Survey (mMOS-SS) in a primary healthcare setting (PHCS). A cross-sectional descriptive study was carried out with Spanish outpatients (N=903) from a PHCS. Subjects responded to the first eight items of theclassic Medical Outcomes Study-Social Support Survey (MOS-SS), in effect the mMOS-SS, and immediately proceeded to give answers to the rest of the scale. Also, sociodemographic, self-reported health status and mood/anxiety disorders data were collected. Factor structure was analyzed using exploratory and confirmatory analysis, internal reliability was determined by Cronbach's standardized alpha, and consistency was evaluated by item-to-total score correlations. Evidences of validity were assessed by Spearman's rank correlations, and using the Mann-Whitney test or Kruskal-Wallis test as appropriate, of differences between the mMOS-SS and the MOS-SS. Mean scores for the MOS-SS (mean=4.26, SD=0.78) and mMOS-SS (mean=4.34, SD=0.74) are similar. Cronbach's alpha was 0.96 and 0.91, for the MOS-SS and mMOS-SS, respectively. Item-to-total score correlations of the MOS-SS and mMOS-SS are ≥0.64 and ≥0.70, respectively. Exploratory factor analysis (EFA) [56.63% variance, KMO=0.904; χ2=4396.27], and confirmatory factor analysis (CFA) [CFI=0.95; NNFI(TLI)=0.97; SRMR=0.05; χ2=296.81; RMSEA=0.17] showed a one-factor structure. Psychometric properties of the Spanish version of the mMOS-SS in outpatients attending a PHCS are adequate. Factor analysis confirmed a one-factor model with acceptable fit indices.