Although the SAQ has been translated into several languages and has been administrated in the United States, United Kingdoms, and a few European countries [17, 19], this is the first time it was translated into Chinese and used in Taiwan. This large study presents the psychometric properties and cross-cultural capabilities of the SAQ scheme. The internal consistency of the SAQ-C is as robust as that of the original English version [16, 17] and Norwegian versions . The SAQ-C scheme is also valid based on its good model construct and significant associations with several safety behaviors.
To date, this is the largest safety culture survey carried out at Taiwanese hospitals. The success of the survey is attributable to several factors. First, the SAQ-C was easily answered and there was a high response rate of 69.4%, which is compatible with the international benchmark of 66-72% . Second, the TJCHA fully supported the survey. The TJCHA is the major force advocating patient safety and medical care quality improvement in Taiwan. Hospital leaders were thus more willing to participate in the survey. Third, all respondents were anonymous, and thus they might feel more comfortable to fill out the questionnaire. Fourth, all participating hospitals benefited from the survey as they were provided with feedback information from the survey. Hospitals received not only their own safety culture information, but also the benchmarking data from 200 hospitals, nearly 40% of the total number of hospitals in Taiwan.
Although many patient safety programs have been initiated in Taiwan, these activities mainly focused on technical and engineering solutions to unsafe working procedures and care processes. Nevertheless, many safety efforts would not be implemented and internalized without changes to organizational culture. This study found that safety culture was not sufficiently established at most Taiwanese hospitals. The mean percentages of positive attitudes toward the five safety dimensions were below the international standard (60%). Furthermore, significantly wide variations in all safety dimensions were also noted, which implied that although a few hospitals had already developed a positive safety culture, more hospitals still lagged behind the population mean. The unsatisfactory survey results are warning signals to healthcare authorities, hospital managers, and the public.
Accumulating evidence supports the relationship between mature safety culture and patient safety, and improving a healthcare organization's safety culture is associated with improved patient outcomes [4, 23–25]. Therefore, the Joint Commission on the Accreditation of Healthcare Organizations in the United States and the National Patient Safety Association in the United Kingdom suggested hospitals should conduct safety culture surveys for safety improvement on a regular basis. Many U.S. hospitals have utilized valid questionnaires to measure safety attitudes among clinical areas [26, 27], and to compare changes in safety attitudes after evidence-based interventions [23, 25, 28]. However, the combination of safety culture changes and safety initiatives is still rare at Taiwanese hospitals.
There is evidently considerable room for improvement in developing a more mature safety culture than the status quo for most Taiwanese hospitals. This study has shown that the SAQ-C is a valid and easily administrated instrument. For the first step, hospitals can use this tool to measure their employees' safety attitudes on a regular basis. Hospital safety managers can track the trends of culture changes of specific clinical units or the hospital as a whole. Moreover, this study shows that there is strong association between safety culture and healthcare workers' safety behaviors (collaboration, safety training, and adverse events reporting), which are closely linked to patient safety. The regular safety survey at clinical units and hospitals can also be used as leading indicators to reflect to some extent the safety index of the clinical area and the likelihood of adverse events. Hospitals then need to review and take appropriate actions in response to unsatisfactory items based on healthcare workers' collective perceptions of their working areas. For instance, the Comprehensive Unit-based Safety Program was developed and introduced to improve the teamwork climate of intensive care units in Michigan, United States . Accumulating reports have demonstrated the improvement in teamwork can significantly improve patient outcomes and decrease avoidable errors [24, 25, 28]. More studies are needed in Taiwan to clarify the causal relationship between safety culture changes and clinical outcomes in different culture settings.
The survey results may serve as a reference in formulating national patient safety policies. For instance, less than one-third of the hospital employees felt positive to their working conditions and the level of staffing got the lowest score (32.8%) among all core items. If staff insufficiency was related to hospital's cost containment strategy and the payment scheme of the National Health Insurance, the healthcare authorities could address this safety concern and put it on the agenda of the annual National Patient Safety Forum led by the Department of Health.
Regular hospital safety culture survey after this study has been fully supported by the TJCHA. A SAQ-C website is initiated http://psc.tjcha.org.tw and hospitals can cooperate with the TJCHA to conduct a unit-wide or hospital-wide survey online. The TJCHA taskforce will help in analyzing the survey data and provide feedback to participating hospitals. Continuous efforts have been initiated to minimize the variations in safety culture among hospitals. After the survey, the TJCHA invited hospitals with high SAQ-C scores to share their successful experiences in establishing safety culture with all participating hospitals. Team resource management programs and evidence-based clinical interventions will be developed and introduced by the TJCHA to help hospitals and their healthcare workers in building a safer environment and culture than the status quo, especially for the hospitals with low SAQ-C scores.