Department of Nutrition, School of Public Health, University of São Paulo, Avenida Doutor Arnaldo, 715, São Paulo, SP, 01246904, Brazil.
Research Institute, Hospital do Coração (HCor), Rua Abílio Sorares, 250, 12° andar, São Paulo, SP, 04005-000, Brazil.
Nutr J. 2018 May 4;17(1):49. doi: 10.1186/s12937-018-0359-5.
The diet of the Brazilian Cardioprotective Nutritional Program (BALANCE) classifies food into four groups and sets the daily amount to be consumed. The dietary approach of BALANCE is different from other dietary recommendations; therefore, it is not possible to use existing dietary indexes (DI) to assess patient's adequacy to BALANCE diet. For this reason, it is important to develop a specific dietary index based on BALANCE diet. This study aims to describe the development of the BALANCE DI, evaluate its internal consistency, construct and content validity and population characteristics associated with the index.
We analyzed baseline data from the BALANCE randomized clinical trial ( https://www.clinicaltrials.gov/ ; NCT01620398). The four food groups of the diet were adopted as index components. Points ranging from 0 to 10 were given to each index component. Internal consistency was evaluated by correlation coefficients between total score and component scores, as well as Cronbach's Alpha. Content and construct validity were assessed by checking how nutrients are associated with the index and if the index could distinguish between groups with known differences in diet, respectively. Crude and adjusted linear regression analyses were performed to evaluate population characteristics associated with the index.
The analysis included 2044 subjects (58.6% men). The average of the total index was higher among women (p < 0,05). The components of the index showed low correlations with each other. The correlations between each individual component with the total index were > 0.40. Cronbach's alpha coefficient was 0.66. High scores in the index were inversely associated (p < 0,05) with energy, total fat, monounsaturated fat (MUFA) and cholesterol; they were positively associated (p < 0,05) with carbohydrates and fiber. Hypertensive men and diabetic women had higher scores, while male smokers had lower scores.
The BALANCE DI showed reliability and construct validity similar to other DI. It also detected characteristics of individuals that are associated with higher or lower index scores.
巴西心脏保护营养计划(BALANCE)的饮食将食物分为四类,并设定了每天的摄入量。BALANCE 的饮食方法与其他饮食建议不同;因此,不可能使用现有的饮食指数(DI)来评估患者对 BALANCE 饮食的适宜性。出于这个原因,基于 BALANCE 饮食制定特定的饮食指数非常重要。本研究旨在描述 BALANCE DI 的开发,评估其内部一致性、构建和内容有效性以及与该指数相关的人群特征。
我们分析了 BALANCE 随机临床试验的基线数据(https://www.clinicaltrials.gov/;NCT01620398)。饮食的四个食物组被采用为指数成分。每个指数成分的分数为 0 到 10 分。通过总评分与成分评分之间的相关系数以及 Cronbach's Alpha 评估内部一致性。通过检查营养素与指数的相关性以及指数是否能够区分饮食差异已知的组来评估内容和构建有效性。分别进行了未调整和调整后的线性回归分析,以评估与指数相关的人群特征。
分析包括 2044 名受试者(58.6%为男性)。女性的总指数平均值较高(p<0.05)。指数的成分彼此之间相关性较低。每个单独成分与总指数的相关性均>0.40。Cronbach's alpha 系数为 0.66。指数得分较高与能量、总脂肪、单不饱和脂肪(MUFA)和胆固醇呈负相关(p<0.05);与碳水化合物和纤维呈正相关(p<0.05)。高血压男性和糖尿病女性的得分较高,而男性吸烟者的得分较低。
BALANCE DI 具有与其他 DI 相似的可靠性和构建有效性。它还检测到与较高或较低指数得分相关的个体特征。