Research Article: Validation of the Indonesian version of the Safety Attitudes Questionnaire: A Rasch analysis

Date Published: April 10, 2019

Publisher: Public Library of Science

Author(s): Evi Ningrum, Sue Evans, Sze-Ee Soh, José-Antonio López-Pina.

http://doi.org/10.1371/journal.pone.0215128

Abstract

Safety climate, which provides a snapshot of safety culture, is rarely measured in Indonesian healthcare organisations because there are no validated surveys that can be administered in its native language, Bahasa Indonesia. The objectives of this study were to translate and linguistically adapt the Safety Attitudes Questionnaire into Bahasa Indonesia, and investigate the internal construct validity and reliability of the translated survey.

The Safety Attitudes Questionnaire was translated into Indonesian language through forward and backward translation. The internal construct validity and reliability of the translated survey was assessed using Rasch analysis which examines overall model fit, unidimensionality, response format, targeting, internal consistency reliability and item bias.

A total of 279 nurses (response rate 82%) completed the Indonesian version of the Safety Attitudes Questionnaire. Most respondents were Division 2 registered nurses (n = 209; 75%), female (n = 174; 62%), and aged less than 30 years (n = 187; 67%). All six domains of the Indonesian version of the Safety Attitudes Questionnaire demonstrated unidimensionality (t-test less than 0.05 threshold value). However, suboptimal targeting (ceiling effect) was observed in all domains, and had at least one misfitting item (item fit residual beyond ±2.5) Item bias was also evident in most domains.

This study has translated and validated an Indonesian version of the Safety Attitudes Questionnaire for the first time. Whilst there was general support to sum items to obtain domain scores, further work is required to refine the response options as well as the wording and number of items in this survey to improve its overall measurement properties.

Partial Text

Safety culture is an aspect of organisational culture that refers to the views, perceptions and actions of personnel within an organisation towards safety management and policy [1, 2]. Due to the complexities of measuring culture [3], safety climate surveys are often used to provide a snapshot of the culture of safety within an organisation [4]. There is emerging evidence demonstrating that low levels of safety climate may lead to poor safety outcomes and adverse events, such as increased length of stay [5], higher postoperative mortality rates [6] and higher 30-day readmission rates [7]. Adverse events in Indonesian hospitals have been reported to be as high as 171 cases among 285 admissions in Jakarta [8]. This suggests that there may be issues around safety climate in these hospitals. Thus, measuring safety climate in this setting may be helpful to guide the design and delivery of appropriate quality improvement programmes to improve patient safety.

Rasch analysis was used to examine the internal construct validity of the SAQ-INA in this sample of Indonesian nurses. There is a growing recognition that modern psychometric approaches based on item response theory (e.g. Rasch analysis) has advantages over classical test theory approaches including CFA [24]. Specifically, Rasch analysis allowed the following measurement properties of the SAQ-INA to be assessed:

In this study, we translated the SAQ into the Indonesian language (SAQ-INA) and examined its capacity to adequately assess safety climate in Indonesian health services using Rasch analysis. Our results demonstrate that all six domains were unidimensional, and that it is appropriate to sum individual items to obtain domain scores. However, there were some notable issues with the instrument; notably ceiling effects were observed for all domains and there were misfitting items. Nevertheless, this is the first study to have translated and validated the SAQ into the Indonesian language using Rasch analysis. The strength of the Rasch measurement model is the ability to identify measurement issues of the SAQ-INA such as response thresholds or item bias by comparing the response patterns of individuals to the entire sample [27, 30]. Thus, our findings can be used to inform the refinement of the SAQ-INA in future studies so that it can accurately measure the level of safety climate in Indonesian hospitals.

This is the first study to have translated and examined the internal construct validity and reliability of an Indonesian version of the SAQ using Rasch analysis. Whilst it is appropriate to sum items to obtain scores for individual domains such as teamwork and safety climate, further studies are required to validate the SAQ-INA in larger samples with a refined response format and item wording so that it can be used to reliably measure the level of safety climate in Indonesian hospitals.

 

Source:

http://doi.org/10.1371/journal.pone.0215128

 

Leave a Reply

Your email address will not be published.