FACTOR ANALYSIS OF THE MINNESOTA LEISURE TIME ACTIVITIES QUESTIONNAIRE - BRAZILIAN PORTUGUESE

Abstract

ABSTRACT Introduction: Accurate measures of physical activity to establish dose-response relationship in health outcomes are still controversial. Scales that estimate caloric expenditure are proposed with a view to categorizing the phenomenon globally. Objective: To carry out a factor analysis of the structure of the Minnesota Leisure Time Activities Questionnaire - Brazilian Portuguese, and propose a new adapted version for the Brazilian elderly community. Methods: Participants were elderly female members of the community, without any distinction in terms of ethnic group and/or social class, who frequented extension projects of two higher education institutions. Those with cognitive impairments; dependence on walking aids; musculoskeletal pain in the last two weeks and neurological diseases were excluded from the study. All participants answered the questionnaire with sociodemographic and clinical data and the Minnesota Leisure Time Activities Questionnaire - Brazilian Portuguese version. We determined whether each activity had been undertaken by the participants and the average number of times it was performed over the last two weeks, considering average time, in minutes, on each occasion. Each activity was considered a single item and energy expenditure was calculated according to the authors’ instructions. We used factor analysis with Varimax orthogonal rotation, excluding items whose variance was equal to zero. Correlation between items was determined using the Pearson correlation matrix. Internal consistency between items, before and after the factor analysis, was verified using Cronbach’s α coefficient. 5% significance level. Results: Participants were 220 women (70.8 ± 5.9 years). The factor analysis resulted in 10 components, which explained, in total, 61.87% of the variance of the total score on the scale. Each component was composed of one, two or three aggregate items. Internal consistency by Cronbach’s α coefficient was 0.30. Conclusion: The factor analysis of the structure of the Minnesota Leisure Time Activities Questionnaire - Brazilian Portuguese- has shown a new range with 10 components, which explained, in total, more than 60% of the variance of the total score on the scale, yet with low internal consistency. Level of Evidence III; Study of nonconsecutive patients; without consistently applied reference ‘‘gold’’ standard.