Transcultural adaptation and psychometric study of the French version of the nursing home survey on patient safety culture questionnaire

BACKGROUND: The Nursing Home Survey on Patient Safety Culture (NHSOPS) questionnaire was developed by the Agency for Healthcare Research and Quality (AHRQ), particularly as an intervention to raise staff awareness about patient safety issues. The main objective of the present study was to provide a validated French-language measure of the safety culture (SC) in nursing homes. Thus the aim was i) to carry out a transcultural adaptation into French of the NHSOPS questionnaire, ii) to assess its psychometric properties in a sample of professionals working in French EHPAD facilities and iii) to develop our own tool. METHODS: The study was carried out on volunteering professionals from 61 nursing homes (from January to March 2016). Two phases were conducted: an initial phase involving the translation and cultural adaptation of the questionnaire, and a second phase in which the psychometric properties of the questionnaire were assessed. A Structural Equation Model (SEM) with a maximum likelihood estimation method was used to evaluate the construct validity of the questionnaire. As the fit of the structure was not sufficient, an exploratory factor analysis using a principal axis factoring with an oblique rotation was then performed. Internal consistency was evaluated and we examined test-retest reliability using Intra-class Correlation Coefficients (ICC). RESULTS: During the initial phase, all items were retained and minor adjustments were made. The participation rate by professionals was 58.4%. The exploratory analysis led to the identification of seven dimensions: Teamwork, Staffing, Compliance with procedures, Handoffs, Feedback and communication about incidents, Supervisor expectations and actions promoting resident safety, Overall perceptions of resident safety and Organizational learning. The SEM confirmed the existence of the seven latent dimensions (CFI = 0.946; TLI = 0.933; SRMR = 0.059; RMSEA = 0.061); internal consistency was acceptable. ICC per item ranged from 0.19 to 0.88. CONCLUSIONS: The results from this study were robust on seven dimensions. This French version is the first on Patient SC to have been applied to the medical-social sector caring for dependent elderly people in France. The NHSOPS questionnaire provides the opportunity to broach this subject. A national evaluation campaign should provide the opportunity to confirm or improve this measure. TRIAL REGISTRATION: NCT02908373 (September 21, 2016) «Retrospectively registered». ELECTRONIC SUPPLEMENTARY MATERIAL: The online version of this article (10.1186/s12913-019-4333-5) contains supplementary material, which is available to authorized users.