Avaliação das propriedades de medida da versão brasileira do questionário Mediterranean Diet Scale (MDS-Brasil) em indivíduos com Diabetes Mellitus

Abstract

Introduction: Diabetes mellitus (DM) is a group of metabolic diseases characterized by hyperglycemia resulting from defects in insulin secretion, insulin action or both. The Mediterranean diet is characterized by low consumption of saturated fat, high intake of unsaturated fatty acids (such as olive oil, fish, or nuts, for example) and high intake of antioxidants, being indicated for individuals with DM. The Mediterranean Diet Scale (MDS) is a tool validated in Canada that allows multidisciplinary screening to assess adherence to the Mediterranean food pattern in a self-administered and pictorial way. The MDS was translated and adapted by our team, through course completion work. Objective: To analyze the measurement properties of the Brazilian Portuguese version of the Mediterranean Diet Scale (MDS-Brazil) in individuals with DM in Brazil. Method: This is a cross-sectional methodological study to assess the following measurement properties of MDS-Brazil: internal consistency, floor and ceiling effect, reproducibility, and construct validity. The 130 volunteers residing in the metropolitan region of Belo Horizonte (MG - Brazil) were recruited. Adults were included, diagnosed with DM (type 1 or 2), not hospitalized, literate and without dietary restrictions and / or visual, mental or cognitive impairment that prevented them from seeing, reading and understanding the text and figures of the instrument. The volunteers were randomized into two groups. For the assessment of test-retest reliability (subgroup 1), 65 participants completed the MDS-Brazil in two stages. To assess construct validity, another 65 volunteers (subgroup 2) filled out the food diary on 3 non-consecutive days and completed the MDS-Brazil. Results: In the sample, 57.7% were women; 56.55 ± 14.88 years; 74.6% DM type 2. The factor analysis (KMO = 0.555 and X² = 137.22; p <0.001) extracted five factors, representing 55.44% of the total variance. No floor and ceiling effects were found. Internal consistency was established by Cronbach's alpha of 0.42. Reproducibility was assessed using the ICC coefficient (0.75; 95% confidence interval [CI]: 0.61 to 0.84). No significant associations were found between MDS-Brazil scores and age (rho = 0.034, p = 0.789), time since diagnosis (rho = -0.178, p = 0.016), education (rho = 0.194, p = 1.22 ), per capita monthly income (rho = 0.047, p = 0.719) and total MDS-Brazil score and 3-day food diary responses (rho = 0.155; p = 0.22). The analysis of the association between the MDS-Brazil score (6.46 ± 1.81) and responses from the 3-day food record (4.52 ± 1.40) demonstrated a T-Test value = 7.39, p <0.001. The scores calculated in the MDS classification showed that 9.2% of the participants were classified as having high adherence to the Mediterranean dietary pattern and 25.4% low adherence. Conclusion: The analysis of the measurement properties of the MDS-Brazil version applied to individuals with DM showed substantial reproducibility; no floor and ceiling effects were found. However, in this population, the instrument had low internal consistency and weak correlations of construct validity.