Avaliação das propriedades psicométricas da versão brasileira do "Personality Inventory for DSM-5 (PID-5)" em amostra de pacientes obesos candidatos a cirurgia bariátrica

Abstract

Background: Personality traits are an important construct in the study of obesity, both to understand its influence on this condition’s origins and to evaluate its role in responses to treatments, such as bariatric surgery. The main models used so far, are based on dimensions of normal personality traits, bringing conflicting results. Hence, there is a demand for instruments that can give more accuracy and improve the study of the personality in this population capturing more dysfunctional traits. In a special section for emerging measures, the new edition of the Diagnostic Statistical Manual of Psychiatric Disorders (DSM-5) proposed an instrument to measure personality mal-adaptive traits: the Personality Inventory for DSM-5 (PID-5), whose psychometric properties have been tested in several samples and localities. Despite promising results, PID-5 has not yet been evaluated psychometrically in other clinical populations where the "personality" construct can have an important impact (in etiologic and prognostic ways), such as obese candidates for bariatric surgery. Objective: The aim of this study is to investigate the psychometric properties of the PID-5 in the evaluation of pathological personality traits in obese candidates for bariatric surgery. This constitutes a fundamental step to assess whether the instrument keeps its original technical features allowing its use and correct interpretation in this population. Setting: Centro da Obesidade e Síndrome Metabólica, Pontifícia Universidade Católica do Rio Grande do Sul (COM-PUCRS), Porto Alegre, Brazil Methods: Cross-sectional study involving 312 medical records of patients who applied for bariatric surgery. We collected demographic data and scores from instruments applied in the psychiatric evaluation routine. The reliability of facets and domains was evaluated through internal consistency indices and the unidimensionality of facets was verified through factorial methods. Exploratory Structural Equations Modeling was performed to test the higher order structure of the model. To investigate convergent validity, linear correlations were calculated between PID-5 scales and instrument scores of correlated constructs (depression, anxiety, defensive styles, and quality of life). Given the characteristic of the sample, we also correlated PID-5 with body mass index (BMI) and binge eating symptoms. Data was analyzed by Factor and Mplus softwares. Results: Facets were unidimensional, with the exception of Risk Taking, better explained by two factors. We replicate the higher order structure of 5 factors of the original model, with good congruence coefficients. The internal consistency of the domains was satisfactory, however, some facets showed fragility of internal consistency, mainly when calculated with more robust methods. In general, the domains and facets correlated as expected with correlated constructs. Conclusion: This study found evidences of adequate psychometric properties of PID-5 for a sample of obese candidates for bariatric surgery. Our results indicate that the instrument is appropriate to be applied in this population, consistently evaluating personality domains and facets of the DSM-5 model. More studies in other centers are needed to replicate and reinforce our findings.