Sociodemographic Data Questionnaire: Instrument developed for this investigation that seeks to collect participants’ characteristics data such as: gender, age, marital status and education.

Outcome Questionnaire Short Form (OQ-10.2) (Lambert et al., 2005): The scale consists of 10 items measured on a 5-point Likert-type scale and assesses the psychological distress levels of adults aged 18 years and older with education from the sixth grade of elementary school onwards. It consists of two subscales: well-being and psychological distress. The original instrument has adequate psychometric properties. The Brazilian version used in this study was developed and validated throughout the study, and its properties are described in the discussion on this subject. The publisher of OQ Measures granted us the license to use this instrument for adaptation, translation and investigation of evidence of validity in Brazil.

Depression Anxiety and Stress Scale Short Form (DASS-21) (Vignola & Tucci, 2014): The scale consists of 21 items with a 4-point Likert-type scale. The emotional states of anxiety, stress and depression are assessed. In the present study, the Cronbach’s alphas obtained were, respectively, 0.90 for depression, 0.88 for anxiety and 0.88 for stress.

Satisfaction with Life Scale (SWLS) (Gouveia et al., 2009): This scale aims to measure the life satisfaction index. There are 5 items that assess life satisfaction indexes through a 7-point Likert-type scale. In our investigation, the scale presented a Cronbach’s alpha of 0.90.

Rosenberg Self-Esteem Scale (RSS) (Hutz & Zanon, 2011): This scale aims to identifying patients’ perceptions of themselves. There are 10 items that must be answered using a 3-point Likert-type scale – 6 referring to a positive view of oneself and 4 to a self-deprecating view. In this study Cronbach’s alpha was 0.91.

Participants were contacted and invited to participate in the study when they sought treatment at a psychotherapy educational institution located in a metropolitan city in the state of Rio Grande do Sul, Brazil and using the snowball sampling method (Breakwell et al., 2010) by completing questionnaires online after the investigation program had been announced in the social media. The selection criteria included being an adult over 18 years of age and having completed the sixth grade of elementary school or higher. No exclusion criteria were used. All participants agreed to collaborate with the study and agreed to its terms. The research was approved by the Ethics Committee of the Pontifícia Universidade Católica do Rio Grande do Sul (PUCRS, Pontifical Catholic University of Rio Grande do Sul) under the number CAAE: 82075318.6.0000.5336.

The procedure adopted for the translation and cultural adaptation of the OQ-10.2 was the standardized process recommended by the editorial team of OQ Measures (Wild et al., 2005). The first step consisted in performing two independent translations into Portuguese from the original instrument, performed by two translators whose native language is Portuguese, fluent in English, psychologists, with a masters degree and with previous experience in translating instruments. A synthesis of the two translations was carried out by a third independent translator, who had the same qualifications and who had the assistance of two independent consultants fluent in English and Portuguese, doctors and experts in the area of clinical psychology and psychotherapy. With the synthesis finalized, the instrument was sent to a committee of four judges who were experts in the area of clinical psychology and psychotherapy; two of them had a masters’ degree and two had a doctorate degree; their task was to investigate the semantic equivalence based on the following criteria: clarity of language, pertinence and theoretical relevance.

Semantic adjustments were made to three items based on suggestions made by the experts’ committee. Thus, item 1 changed from “I am happy now” to “I am happy”; item 4 changed from “I feel loved and desired” to “I feel loved and wanted by other people”; and item 7 changed from “I feel that something is wrong with my mind” to “I feel that something is wrong with the way I think”. Finally, the Portuguese version was back-translated into English by a fourth independent translator, a PhD in clinical psychology and psychotherapy, fluent in Portuguese and with extensive command of the English language. The back-translation was sent to the original authors to ensure approval of content equivalence. After approval by the original authors, which occurred without the need for further changes, an operational equivalence was made regarding the way the instrument was applied. Care was taken to ensure that the format of the questionnaire would be similar to the original, as well as the format of administration and instruction to participants.

The next stage consisted of a pilot evaluation with the target audience. The instrument was applied individually to ten adults who commented on each item regarding its comprehensibility and suitability, and did not present any difficulties nor generated any questions with them. The final stage consisted of a formal study with the target population.

In order to investigate evidence of validity of the instrument internal structure, factor analyses and parallel analyses were conducted using the psych package of the R software (Revelle, n.d.). An exploratory factor analysis with oblique rotation (promax) was performed. The number of factors to be interpreted followed the criterion of parallel analysis using parametric methods and permutation of sample values (Timmermann & Lorenzo-Seva, 2011). To calculate the general index of psychological distress, a Principal Component Analysis (PCA) was used, forcing the extraction of a component. The linear associations of each item of the OQ-10.2 with the general index were investigated. This component corresponds to the weighted sum of all items on the scale. The R software was also used to calculate Cronbach’s alphas for the general scale and the two subscales.

Network analysis was used to investigate the pattern of relationships between the OQ-10.2 items (Epskamp & Fried, 2018). A network analysis comprises two steps. First, a matrix of partial correlations between the instrument items, i.e., conditional associations, is estimated. The matrix is regularized in order to set small correlations to zero to avoid overfitting the model. After this step, the variables were represented in a two-dimensional network, so that vertices represent the variables and edges reveal their association. The thickness and pattern of the edges indicate the direction (positive and negative) and the magnitude of the associations. A positioning algorithm is used to bring the variables closer or further apart according to their associations.

In addition to the network analysis, a community analysis was conducted. Communities are subgroups of vertices in a graph, in order to identify sets of variables that are most associated with each other. The algorithm employed uses a hierarchical method and has as a parameter the modularity measure, which represents the level of intragroup association of the communities (Yang et al., 2016). Modularity measures are estimated from the large set model to an independence model, in which the variables alone constitute communities. The solution that optimizes the modularity value is chosen. Centrality measures helped to describe the role of each variable in the system. The closeness measure indicates which node has the shortest distance from the other nodes in the network, with the distance being the inverse of its association. The expected influence measure is based on the product of the first and second degree associations of each node, indicating its potential for propagation in the network, that is, once activated, its tendency to activate (positive values) or deactivate (negative values) neighboring nodes (Robinaug et al., 2016).

Validity evidence was investigated in relation to external variables, estimating bivariate correlations and partial correlations through network analysis. Spearman correlation analyses of the OQ-10.2 – considering the non-normal distribution of some of the instrument’s variables – were also performed with measures of self-esteem, life satisfaction, symptoms of stress, anxiety and depression.