Avaliação psicométrica das escalas: satisfação com o suporte social e avaliação do empowerment em pessoas que vivem com o diabetes no contexto do Letramento em Saúde

Abstract

Introduction:The Quality of Life (QV) of people with Diabetes Mellitus (DM) is related to self-care resulting from social support, Health Literacy (HL) and people's empowerment. Good LS levels can promote healthy choices and effectiveness in controlling DM. Objective: The validity and reliability of the satisfaction with social support scales (ESSS) and Diabetes Empowerment Scale-Short Form (DES-SF) were evaluated, and the effect of the satisfaction with social support constructs, hypothetically mediated by the LS-D, was analyzed. (base: integrated LS model by Sørensen et al., 2012) and through empowerment in the QV (D21) of people with DM, users of units of the Family Health Strategy (ESF) of Primary Health Care in Montes Claros, MG. Methodology: Cross-sectional study (2016 and 2018) among 195 people with DM. Statistics: confirmatory factor analysis, logistic regression techniques and Structural Equation Modeling (SEM). To check the adequacy of the sample (above 60%) the following tests were estimated: Bartlett sphericity (p<0.05) and Kaiser-Meyer-Olkin (KMO). The Diagonally Weighted Least Squares estimation method was used. (WLSMV). Results: In the 15-item ESSS, a four-factor (1st order) and single-factor (2nd order) structure was obtained among diabetics (2016-2018), in which confirmatory factor analysis (good fit, χ² values should not be significant : p>0.05); χ²/gl= ≤ a 5; CFI and TLI= ≥ 0.90; RMSEA must be less than 0.08, with confidence interval (upper limit) <0.10) indicated adequate fit with χ² (gl)=114.78 (84)*; χ²/gl=1.36. The comparative fit index (CFI)=0.97 and the Tucker-Lewis index (TLI)=0.96. The standardized root mean residual (SRMSR)=0.064 and 0.066, the Root Mean Square Error of Approximation (RMSEA) (90% CI)=0.037 (0.017–0.054). The general internal consistency of the instrument was satisfactory: Cronbach's α=0.69 and McDonald's ômega=0.67 and composite reliability= 0.82. In 2nd order: χ²(gl)=119.92 (86)*; χ²/gl=1.39; CFI=0.97; TLI=0.96; SRMR=0.066 and RMSEA (90%CI)=0.039 (0.020–0.055). The DES-SF scale (8 items), χ² values were not significant: χ²=9.759; gl=20; p=0.972 and χ²/gl are within the recommended range: χ²/gl =0.48. The indices of CFI=0.99, TLI=0.99, RMSEA=0.00 and SRMR=0.04 supported the model. Regarding the validity of the hypothetical model through confirmatory factor analysis, which confirmed its factor structure: χ²=0.00; χ² (gl)=2127.00 (1313.00); χ²/gl=1.61; CFI=0.93; TLI=0.93; SRMR=0.11; RMSEA (90% CI)=0.05 (0.05–0.06). To verify the plausibility of the theoretical model, the Structural Equation Modeling (SEM) technique was used, a multivariate analysis based on regression that allows the analysis of multiple dependent, observable or latent variables. To check the adequacy of the sample (above 60%) the Bartlett sphericity test (p<0.05) and the Kaiser-Meyer-Olkin (KMO) test were estimated. The Diagonally Weighted Least estimation method was used. Squares (WLSMV),suitable for ordinal and nominal categorical data. The adjustment indices were: χ², χ²/gl, CFI, TLI, SRMR and RMSEA. The significance level adopted was p<0.05. The results were consistent with previous literature, corroborating the relevance of these constructs in DM management. Conclusions: The study showed thatthe effect of the constructs on the QV of people with DM confirms the plausibility of the evaluated model.