Psychometric properties of the Latino Students Patient Safety Questionnaire, Brazilian version.

OBJECTIVES: to assess the psychometric properties of the Latino Students Patient Safety Questionnaire, Brazilian version. METHODS: a methodological study, carried out between April 2020 and January 2021, with 218 nursing and medicine students. Structural and discriminant construct validity were assessed by confirmatory factor analysis and cross factor loadings. Reliability was verified by McDonald's omega, average variance extracted, composite reliability, and item-total correlation. RESULTS: the final model reproduced the original structure of 21 items, distributed in five dimensions, requiring the exclusion of one item. Acceptable fit indexes were obtained (x2/gl=2.325; CFI=0.99; TLI=0.98; RMSEA=0.054). Discriminant validity was confirmed. Reliability indicators were adequate, except McDonald's omega, in one factor (0.68), and average variance extracted, in two factors (0.41; 0.47). CONCLUSIONS: the instrument demonstrated evidence of internal validity and satisfactory reliability among nursing and medical students.

Psychometric properties of the Latino Students Patient Safety Questionnaire, Brazilian version / Propiedades psicométricas del Latino Students Patient Safety Questionnaire, versión brasileña / Propriedades psicométricas da versão brasileira do Latino Students Patient Safety Questionnaire

ABSTRACT Objectives: to assess the psychometric properties of the Latino Students Patient Safety Questionnaire, Brazilian version. Methods: a methodological study, carried out between April 2020 and January 2021, with 218 nursing and medicine students. Structural and discriminant construct validity were assessed by confirmatory factor analysis and cross factor loadings. Reliability was verified by McDonald's omega, average variance extracted, composite reliability, and item-total correlation. Results: the final model reproduced the original structure of 21 items, distributed in five dimensions, requiring the exclusion of one item. Acceptable fit indexes were obtained (x2/gl=2.325; CFI=0.99; TLI=0.98; RMSEA=0.054). Discriminant validity was confirmed. Reliability indicators were adequate, except McDonald's omega, in one factor (0.68), and average variance extracted, in two factors (0.41; 0.47). Conclusions: the instrument demonstrated evidence of internal validity and satisfactory reliability among nursing and medical students.

RESUMEN Objetivos: evaluar las propiedades psicométricas de la versión brasilena Latino Students Patient Safety Questionnaire, versión brasilena. Métodos: estudio metodológico, realizado entre abril de 2020 y enero de 2021, con 218 estudiantes de pregrado en enfermería y medicina. La validez de constructo estructural y discriminante se evaluó mediante análisis factorial confirmatoria y cargas factoriales cruzadas. La confiabilidad fue verificada por el omega de McDonald, Ia varianza promedio extraída, Ia confiabilidad compuesta y la correlación ítem-total. Resultados: ei modelo final reprodujo Ia estructura original de 21 items, distribuídos en cinco dimensiones, requiriendo la exclusión de un ítem. Se obtuvieron índices de ajuste aceptables (x2/gl=2,325; CFI=0,99; TLI=0,98; RMSEA=0,054). Se confirmó la validez discriminante. Los indicadores de fiabilidad fueron adecuados, excepto el omega de McDonald, en un factor (0,68), y la varianza media extraída, en dos factores (0,41; 0,47). Conclusiones: el instrumento demostró evidencias de validez interna y confiabilidad satisfactoria entre estudiantes de enfermería y medicina.

RESUMO Objetivos: avaliar as propriedades psicométricas da versão brasileira do Latino Students Patient Safety Questionnaire. Métodos: estudo metodológico, realizado entre abril de 2020 e janeiro de 2021, com 218 estudantes de graduação em enfermagem e medicina. Validades de construto estrutural e discriminante foram aferidas pela análise fatorial confirmatória e cargas fatoriais cruzadas. A confiabilidade foi verificada pelo ômega de McDonald, variância média extraída, confiabilidade composta e correlação item-total. Resultados: o modelo final reproduziu a estrutura original de 21 itens, distribuídos em cinco dimensões, sendo necessária a exclusão de um item. Obtiveram-se índices de ajustes aceitáveis (x2/gl=2,325; CFI=0,99; TLI=0,98; RMSEA=0,054). Confirmou-se a validade discriminante. Os indicadores de confiabilidade foram adequados, exceto ômega de McDonald, em um fator (0,68), e variância média extraída, em dois fatores (0,41 ; 0,47). Conclusões: o instrumento demonstrou evidência de validade interna e confiabilidade satisfatória entre estudantes de enfermagem e medicina.