Fantastic Lifestyle Questionnaire applied to undergraduate medical students during the COVID-19 pandemic: a factor analysis

OBJECTIVE: To identify what structure represents life style in medical students, in a public university, with the application of the Fantastic Lifestyle Questionnaire. METHODS: This is a cross-sectional exploratory study. The Fantastic Lifestyle Questionnaire was applied online to medical students. Factorial analysis was realized and factors were extracted by principal components method. Rotation was conducted by Varimax method. RESULTS: Sixty-one questionnaires were analyzed. Students were male in 45.9%, single in 88.5%. The model was composed of three factors that explained 63.5% of all observed variance. The factors related to sleep perception, nutrition and stress, followed by relationships and entertainment and then, activity. Cronbach’s alpha was 0.81, which was considered good. CONCLUSION: The model composed by the three factors observed in this group of students represented the construct quality of life, evaluated by the Fantastic Lifestyle Questionnaire. This result may provide substrate to actions that aim to improve quality of life and well-being in medical students from this university.

prior and informed consent document, and were fully informed about the objectives and purposes of the study, in addition to the guarantee of anonymity, as well as the guarantee of access to an electronic copy of the consent document. The answers were registered in the research administration application, which has security mechanisms, such as the non-identification of students in the form, restriction of access to answers through e-mail and password access by researchers.
The instrument used to measure the students' lifestyle was the Brazilian version of the FLS questionnaire. This questionnaire consists of nine domains: family and friends, physical activity, nutrition, smoking and drugs, alcohol, sleep, stress and safe sex, type of behavior, introspection and work. Data regarding gender, age, skin color and marital status of the participants were also collected.
Statistical analysis was performed using factor analysis, a multivariate method that provides for the reduction of a group of variables into a smaller set of factors that represent the underlying latent structural dimensions. The methodological basis of factor analysis resides in the assumption that some underlying factors, in smaller numbers than the observed variables, are responsible for the covariation between the set of variables. In this way, it is possible to identify constructs that summarize or explain the set of variables observed. For this purpose, the adequacy of the database, the extraction technique and the type of factor rotation must be observed 10 .
For the analyses, the Stata program version 13.0 (Statacorp, L.C.) was used. Qualitative variables are presented in percentages. For the factor analysis, the correlations between the variables were initially observed, and correlations with values greater than 0.3 in the matrix were considered adequate and the others excluded. Of the 19 variables, 10 remained in the model. This number of variables in the model in relation to the sample size resulted in a 1:5 ratio, considered appropriate. Bartlett's test of sphericity was applied, which was significant (p<0.001). Subsequently, the Kaiser-Meyer-Olkin (KMO) test was applied on the selected matrix, to show the proportion of the common variance. This initial analysis revealed that factorization would be adequate.
The extraction of factors was performed using the principal components model, to estimate the best combination of variables that would explain most of the observed variance. The factors were selected by the latent root criterion (Eigenvalue), keeping those with eigenvalues greater than 1.0. The Eigenvalues were observed by the Scree graph. The rotation of the factors to adjust the loads was performed using the Varimax method. The variables were related to the factors according to their highest rotated loads and the factors were named according to the variables with the greatest weight in their construction. For the construction of this model, parsimony was used. Variances, commonalities and factor loadings are presented.

RESULTS
The sample consisted of 61 participants and their characteristics are shown in Table 1. Regarding the global assessment of the FLS questionnaire, one participant (1.63%) had a score between 70 and 84-very good; 34 (55.73%) between 55 and 69-good; 24 (39.34%) between 35-54-regular, and two (3.27%) between 0 and 34-needs improvement.
Ten variables participated in the factor analysis: 1. I have someone to talk to about the things that are important to me; 2. I give and receive affection; 3. I am vigorously active for at least 30 minutes a day; .001) indicated that the correlational matrix was not an identity matrix, which was followed by a factor analysis.
The analysis of eigenvalues suggested the existence of three main factors (Table 2), which can be confirmed by the Scree graph ( Figure 1). The factor analysis model identified three factors that were named taking into account the factor loadings obtained. Thus, Factor 1 refers to the perception of sleep, diet and stress, Factor 2 refers to the perception of relationships and leisure and Factor 3 refers to the perception of activity (Figure 1). These three factors accounted for the structure found in the analyzed questionnaires.
The three factors identified explained 63.50% of the total variance observed. Varimax rotation was performed, and the summary of the results of the factor analysis is also found in Table 2. Table 2. Results of the analysis of factor eigenvalues, proportion of explained variance and accumulated variance factor followed by the factor solution with rotated loads, commonalities, percentage of explained variance of the Fantastic Lifestyle Questionnaire applied to medical students.

DISCUSSION
This study involving medical students from a public university analyzed the multidimensional characteristics of the quality of life construct. The model resulted in three factors that explained 63.40% of all observed variance. These factors were related to the perception of sleep, food and stress, relationships, leisure and activity. The internal consistency of the FLS questionnaire, in the context analyzed, measured by Cronbach's alpha was 0.81. This internal consistency denotes the degree of internal relationship between the items in the questionnaire and the consistency of the responses of the studied group. It has been recommended over the years that the cut-off point for the value of this statistic should be 0.70 11 . Thus, the value reached in this study was good, revealing consistency of information.
The components of food, sleep and stress self-control were correlated and revealed great importance in this study. This result suggests that these variables were very important to determine the quality of life in this group of individuals. High factor loadings were observed for the variables "being in the range of weight considered healthy" and "sleeping well and feeling rested". This relationship has also been described in the literature. Vernia et al. discussed the complex interrelationship that exists between sleep disorders and their relationship to poor eating habits. In fact, these disorders are considered "triggers" for digestive disorders. On the other hand, individuals who eat poorly generally sleep less and not well. All these circumstances affect the quality of life and deserve attention 12 .
The second factor identified in medical students was formed by the components of relationships, affection and leisure. It is  essential to rescue the importance of this component for the construction of quality of life, especially during the COVID-19 pandemic, where people around the world underwent a process of reinvention and needed to give new meaning to their interpersonal relationships. Self-care, represented by leisure activities, was also part of this factor. It has already been pointed out that, in medical students, resilience is positively correlated with quality of life, whereas psychological stress is negatively correlated 13 . Also, self-care and work-life balance have been identified as essential in medical students not only for quality of life, but for the delivery of effective care. Medical professors must be proactive in supporting these practices in students, particularly in critical periods, such as the COVID-19 pandemic 14,15 .
In the present study, the participant's daily activity dimension constituted the third factor, which refers to domain A of activity, which makes up the name FANTASTIC. In a Brazilian multicenter study that included twenty-two medical schools, it was observed that 40% of students reported not having time for physical activity. Furthermore, an association between quality of life and physical activity was revealed, including a dose-effect relationship (the greater the activity, the better the quality of life) 16 .
It is interesting to identify that the domains related to the use of cigarettes and drugs, alcohol and career (which represent domain C of FANTASTIC) did not participate in the making of this construct. It is worth remembering that this particularity refers to this group of individuals, as it is known that the consumption of these substances is negatively associated with quality of life. Among nursing students in Australia, alcohol consumption, physical inactivity and skipping breakfast were associated with poor quality of life 17 . In the present study, career or work were not very important, possibly because they were students.
This study has limitations. First, it was carried out during the COVID-19 pandemic and, therefore, the forms were filled out online, voluntarily. The sample refers to a single public university, making it difficult to generalize the results. The small sample size may have influenced the results. However, given the lack of studies on this topic, especially in Brazil, this study contributes to the reflection on the lifestyle of medical students. Future studies with a larger sample size and after the COVID-19 pandemic may contribute to a greater understanding of the topic.

CONCLUSION
This study evaluated the components of the FLS in order to identify, in this group of medical students, which elements shape its construct. Thus, components related to the perception of sleep, food and stress, relationships, leisure and activity were identified. This result can provide substrate for actions to improve the quality of life and well-being of students at this university.