Statistical analysis

We conducted descriptive statistics to examine the underlying assumptions of normality for all variables of interest. The Cronbach’s alpha and Spearman-Brown were used for the assessment of the MQOL-R’s reliability. For the BrP-MQOL-R, the maximum likelihood factor analyses with oblique rotation were conducted. We checked the scale's internal structure using confirmatory factor analysis (CFA) and establishing its reliability and validity. Items with a loading equal to or higher than 0.4 were retained to be considered relevant [17]. Factors that win with eigenvalues greater than one were also excluded. Convergent validity was evaluated by Pearson’s correlation coefficient between BrP-MQOL-R total scores, subscales, and the SIS measuring overall quality of life with scores on NPS (0–10) and the KPS scale. The non-parametric receiver operating characteristics (ROC) analyses, with the exact binomial of the area under the curve (AUCs) with 95% confidence intervals (CI), is presented. We calculated the standard errors (SEs) by Hanley’s method [18]. The cutoff values with the highest Youden index, with 90% sensitivity and 100% specificity, are presented for the BrP-MQOL-R with a ROC AUC 0.70. Finally, a stratified-by-sex analysis was used to assess the correlation between age, education level, if they were hospitalized when they answered the MQOL-R (Yes/No), and the scores of the dependent variable MQOL-R. We employed regression analysis with a stepwise forward technique. The prior sample size was estimated a priori based on the number of volunteers' ratio to the number of items. In this case, the MQOL-R has 14 questions. Based on this criterion, we needed 140 volunteers. Considering potential loss by insufficient data, we increased the sample size by 10% [10]. For all statistical analyses, significance was set at P < 0.05. The analysis used SPSS version 24.0 (IBM, Armonk, NY, USA), and the CFA was conducted by means of SPSS. AMOS. Version 24.0 (IBM, Armonk, NY, USA).