Comparison and convergent validity of five Mediterranean dietary indexes applied to Brazilian adults and older adults: data from a population-based study (2015 ISA-Nutrition)

Abstract

Different dietary indexes are proposed to investigate adherence to the Mediterranean diet (MD). However, they are based on different methodologies, and limited research has compared them to each other, particularly in non-Mediterranean populations. We aimed to compare five indexes intended to measure adherence to the MD. The sample was composed of adults and older adults (n 1187) from 2015 ISA-Nutrition, a cross-sectional population-based study in São Paulo, SP, Brazil. Dietary data obtained through two 24-h dietary recalls (24HDR) from which the Mediterranean diet scale (MDS), Mediterranean diet Score (MedDietscore), Mediterranean dietary pattern (MDP), Mediterranean Adequacy Index (MAI) and Mediterranean-Style Dietary Pattern Score (MSDPS) were calculated. The correlations and agreements between them were analysed by Spearman's correlation and linearly weighted Cohen's Kappa coefficients, respectively. Confirmatory factor analyses (CFAs) were applied to investigate their convergent validity. The highest correlations were found between MDP and MAI (r = 0⋅76; 95% CI 0⋅74-0⋅79) and between MDP and MDS (r = 0⋅72; 95% CI 0⋅69-0⋅75). The greatest agreements observed were moderate, between MDP v. MAI (κ = 0⋅57, P < 0⋅001) and MDP v. MDS (κ = 0⋅48, P < 0⋅001). The goodness-of-fit of CFA for MedDietscore (RMSEA = 0⋅033, 90% CI 0⋅02-0⋅042; SRMR = 0⋅042) and MSDPS (RMSEA = 0⋅028, 90% CI 0⋅019-0⋅037; SRMR = 0⋅031) had acceptable values for absolute fit indices. Vegetables, olive oil, MUFA:SFA ratio and cereals with legumes were more relevant to characterise the MD (factor loadings ≥0⋅50). The MDS, MAI and MDP classified the population similarly, but the MedDietscore showed better performances in evaluating adherence to the MD. These results provided guidance for the most appropriate Mediterranean dietary index to be applied in non-Mediterranean populations.

Keywords: Adherence; Confirmatory factor analysis; Dietary indexes; Mediterranean dietary indexes; Mediterranean dietary pattern.

Fig. 1.

Confirmatory factor analysis for each Mediterranean dietary index. Circles: latent variables. Squares: observed variables. The straight arrows connect the observed and the latent variables: standardised factor loadings. The free straight arrows attached to observed variables indicate measurement errors. (a) MedDietscore’ construct: potato (pot), vegetables (veg), legumes (leg), poultry (poul), olive oil (olioil); (b) MSDPS’ construct: whole grains (grains), fruits, poultry (poul), dairy, potato (pot), olives, legumes and nuts (ollegnu), olive oil and vegetables (oilveg); (c) MDS’ construct: fruits and nuts (frunut), vegetables (veg), legumes (leg), MUFA:SFA (mufsfa); (d) MDP’ construct: vegetables (veg), legumes (leg), MUFA:SFA (mufsfa), trans-fatty (trans), alcohol (alco); (e) MAI’ construct: one (carbohydrate group), two (protective group), three (land animal group), four (sweets).