Inconsistencies Between Two Cross-Cultural Adaptations of the Hospital Survey on Patient Safety Culture Into French

Abstract

Objectives: Two cross-cultural adaptations of the 12-dimension Hospital Survey on Patient Safety Culture (HSOPSC) into French coexist: the Occelli and Vlayen versions. The objective of this study was to assess the psychometric properties of the Occelli version in comparison with those reported for the Vlayen and the original US versions of this instrument.

Methods: Using the original data from a cross-sectional study of 5,064 employees at a single university hospital in France, we examined the acceptability, internal consistency, factorial structure, and construct validity of the Occelli version of the HSOPSC.

Results: The response rate was 76.8% (n = 3888). Our study yielded lower missing value rates (median, 0.4% [range, 0.0%-2.4%] versus 0.8% [range, 0.2%-11.4%]) and lower dimension scores (median, 3.19 [range, 2.67-3.54] versus 3.42 [range, 2.92-3.96]) than those reported for the Vlayen version. Cronbach alphas (median, 0.64; range, 0.56-0.84) compared unfavorably with those reported for the Vlayen (median, 0.73; range, 0.57-0.86) and original US (median, 0.78; range, 0.63-0.84) versions. The results of the confirmatory factor analysis were consistent between the Vlayen and Occelli versions, making it possible to conduct surveys from the 12-dimensional structure with both versions.

Conclusions: The inconsistencies observed between the Occelli and Vlayen versions of the HSOPSC may reflect either differences between the translations or heterogeneity in the study population and context. Current evidence does not clearly support the use of one version over the other. The two cross-cultural adaptations of the HSOPSC can be used interchangeably in French-speaking countries.