Factor Structure and Reliability of MOS Social Support Survey among Mexican Women with Breast Cancer

Abstract

Background: Social Support (SS) has been an important variable to consider in the psychosocial understanding of patients with cancer. Objective: To test the 3-factor model for the Medical Outcomes Study Social Support Survey (MOS SSS) and estimate its internal consistency reliability in Mexican women with Breast Cancer (BC). Methods: 300 women diagnosed with BC were enrolled in a convenience sampling. Participants were asked to complete the Spanish version of the MOS SSS (Argentina). After the approval of the Ethics and Research Review Board of Instituto Nacional de Cancerología, the Mexican adaptation was created through three phases: 1) linguistically adaptation; 2) presentation of the corrected and integrated version to another expert, and 3) pilot study. Results: The MOS SSS obtained a factorial structure with 13 items distributed among three factors. A high internal consistency (Alpha=0.912) and an acceptable explained-variance value (67%) were obtained; finally, the confirmatory analysis determined a proper adjustment with that structure. Conclusions: This instrument has adequate psychometric properties and is useful for evaluation and culturally appropriate to the clinical management of SS.

The Medical Outcome Study Social Support Survey (MOS‐SSS): A psychometric systematic review

“…Four studies claimed convergent validity by assessing the correlations between items and their subscales or correlations between the subscales, which is not in line with the COSMIN guidelines (Din et al, 2020; Dumitrache et al, 2021; Gálvez‐Hernández et al, 2020; Norhayati et al, 2015). Their reports of this “convergent validity” were not included in the review.…”

“…Thirty‐four studies performed structural validity tests. Six studies performed on both EFA and CFA (Bavarsad et al, 2021; Gálvez‐Hernández et al, 2020; Gómez‐Campelo et al, 2014; Moser et al, 2012; Sherbourne & Stewart, 1991; Zucoloto et al, 2019). Twelve studies performed EFA only, of which two performed Rasch analysis (Kim et al, 2017; Kim & Mazza, 2014).…”

“…Of 18 studies performing EFA, nine studies used principal component analysis (PCA), two studies used principal axis factoring (PAF; Gálvez‐Hernández et al, 2020; Mahmud et al, 2004), two studies used maximum likelihood principle (Bavarsad et al, 2021; Zucoloto et al, 2019) and five studies did not report a specific extraction principle (Gómez‐Campelo et al, 2014; Kim et al, 2017; Saddki et al, 2017; Yilmaz & Bozo, 2019; Yu et al, 2015). Ten studies used the orthogonal rotation method (varimax (9), unspecific (1) (Nicolaou et al, 2015)).…”

“…Ten studies used the orthogonal rotation method (varimax (9), unspecific (1) (Nicolaou et al, 2015)). Three studies used the oblique rotation method (oblimin (2) (Gálvez‐Hernández et al, 2020; Mahmud et al, 2004), unspecific (1) (Yu et al, 2015)). One stated unrotated (Sherbourne & Stewart, 1991), and four studies did not report a specific rotation method (Gómez‐Campelo et al, 2014; Kim et al, 2017; Saddki et al, 2017; Yilmaz & Bozo, 2019).…”

“…One study claimed discriminant validity by comparing the correlation between an item and its two standard errors in their own subscale and the correlation between an item and its two standard errors in any other subscale (Dafaalla et al, 2016). Four compared correlations between factors and Average Variance Extracted (AVE) square root for discriminant validity (Din et al, 2020; Dumitrache et al, 2021; Gálvez‐Hernández et al, 2020; Norhayati et al, 2015). These approaches were not in line with COSMIN guidelines.…”