The safety attitudes questionnaire – ambulatory version: psychometric properties of the Slovenian version for the out-of-hours primary care setting

Patient safety is an important aspect of quality assurance and improvement in health care. In recent years, a concept of patient safety culture has been developed. It describes leader and staff interaction, attitudes, routines, awareness, and practices, which impinge on the risk of patient adverse events [1]. Safety culture is regarded as a group phenomenon rather than that of individuals [2]. In organizational psychology research, safety culture is described by both qualitative and quantitative methods [3]. Quantitative surveys have concentrated on measuring staff perceptions, which are referred to as organizational climates. Organizational climates are mathematical expressions of how members in natural social units perceive that cultural norms are enacted by leadership and members in the unit. These climates are measured both according to level of mean and the degree to which staff share the perceptions, which is the organizational climate strength [4]. Organizational climates with diverging perceptions amongst staff are regarded as weak with limited power to predict staff practices [5].

Primary health care differs from hospitals in terms of organisational structure, administrative and clinical processes and the reasons for encounters [6]. Therefore, also patient safety culture dimensions could differ when compared to secondary or tertiary health care.

So far, several studies have addressed the issue of patient safety culture in primary health care [7‐10]. The revealed dimensions were similar but not identical. The most comprehensive view was provided by Kirk et al. [10] who conducted a literature review of patient safety dimensions and put forward nine dimensions that patient safety culture in primary care was expressed by: overall commitment to quality, priority given to patient safety, perceptions of the causes of patient safety incidents and their identification, investigating patient safety incidents, organisational learning following a patient safety incident, communication about safety issues, personnel management and safety issues, staff education and training about safety issues, and teamworking around safety issues.

Out-of-hours care (OOHC) services are primary health care services enabling access to primary health care services during out-of-hours working hours (weeknights and weekends) [11]. There are many different models for organizing OOHC services [11] and yet no clear definition of OOHC has been developed in the literature [12].

In Slovenia, OOHC and emergency medical services (EMS) are combined and are available both at the same place and time. Professionals that work in OOHC are family physicians, emergency physicians, and emergency nurses. Sometimes, there are also laboratory technicians and radiology technicians. Usually, emergency nurses work only in OOHC settings while family physicians work in their practice and in OOHC on the basis of rotation. One or two teams are on call at the same time on 8-hours rotation. An OOHC team consists of a family- or emergency physician and two emergency nurses. These teams are located in primary health care centres and available 24/7 enabling free access of patients on their own demand. If there is an on-the-field emergency, the team leaves the OOHC surgery in order to respond to the call. This is a major difference between Slovenian and most European OOHC centres. In a majority of European countries OOHC centres are organized separately from EMS [11]. An OOHC centre in Slovenia has thus a unified leadership and can be seen as a “natural social unit”; which is a validation criteria for organizational climate measurements [2]. Patient safety in OOHC has been studied only in few European countries [9, 13‐16]. These studies dealt with several aspects of patient safety, i.e. patient safety incidents [15] and triage [13].

Several tools have been developed to measure safety attitudes of health care providers [7, 17‐22]. The Safety Attitudes Questionnaire (SAQ) is one of the widely used and also one of the most appropriate instruments for measuring patient safety culture [22‐24]. It measures six factors: Teamwork climate, Safety climate, Job satisfaction, Perceptions of management, Working conditions and Stress recognition [25]. In 2007, the SAQ was adapted to outpatient (primary health care) settings [7]. This SAQ – Ambulatory Version (SAQ-AV) proved to be a reliable tool for comparing attitudes across different professional groups of health care providers outside hospitals [7]. In 2014, Bondevik et al. [16] tested the Norwegian version of SAQ-AV and proved it to be a reliable tool for measuring patient safety culture in OOHC in Norway.

In Slovenia, patient safety culture in OOHC has not been studied and no tools for measuring it have yet been developed or adapted. The purpose of this study was to translate the SAQ-AV to Slovenian language and to adapt it to the Slovenian OOHC settings. We had the following aims of the study: 1) to translate the English version of SAQ-AV to Slovenian language; 2) to test the reliability of the Slovenian version of SAQ-AV in OOHC settings; and 3) to determine the factor structure of the Slovenian version of SAQ-AV in OOHC settings.

The Slovenian version of SAQ-AV, EFA revealed five factors, some of which were similar, but not identical to previous versions of the SAQ-AV [7, 16]. This study was a part of the international study lead by the Norwegian researchers who had adapted the SAQ-AV for Norwegian primary care settings. The Norwegian factor structure [16] provided five factors: Teamwork climate, Safety climate, Job satisfaction, Working conditions, and Perception of management. The original SAQ, developed at the University of Texas at Austin [25], described six factors: Teamwork climate, Safety climate, Working conditions, Job satisfaction, Perceptions of management and Stress recognition.

Two of the factors in our study had Cronbach’s alphas lower than the recommended 0.7 [28]. Considering cultural differences between the countries this is not surprising. Although content of four factors had similarities in the Norwegian and Slovenian factor structure, several items loaded on different factors in Slovenia compared to Norway.

This indicates that the perceptions of safety culture in Norway and Slovenia may be different. Items related to perceived problems voicing concerns regarding patient safety, emerged as an independent factor in our study. We named this factor Communication. Items included in this factor in Slovenia belong to several factors found in other studies [25, 29]: Teamwork climate, Safety climate, and Perceptions of management. Actually, Communication is a part of all these fields but it seems that it was so important to the Slovenian participants that it emerged as an individual factor.

There is some overlap between Communication and the concept of psychological safety. The term psychological safety refers to which extent team members feel comfortable seeking feedback, sharing information, asking for help, talking about errors, and experimenting [30‐32]. Psychological safety is a requirement for effective communication regarding risk in patient care [31], and is complementary to a good safety culture [33]. Psychological safety has been recognised as safety relevant in other health care settings [30, 33] but not in previous analyses of SAQ-AV data nor in OOHC settings. Our results may indicate that Communication and psychological safety are perceived more important and perhaps cultural traits more at stake to participants in Slovenia compared to other countries, since they possibly emerge as individual factors without the study seeking to map them explicitly. The finding supports that Communication and Psychological safety are considered possible independent factors in future safety culture surveys.

Our study has some methodological limitations which need to be addressed. The first one is the process of factor analysis. We chose not to proceed by confirmatory factor analysis despite the fact that it is usually a key step in psychometric evaluation of survey tools. Our decision was based on small sample size which was a consequence of specific study settings.

The second one is the low value of Cronbach’s alpha in two of the factors. These are the same factors that have only three items. As the alpha value is related to the number of items in the scale [34], the low values of Cronbach’s alpha could be a consequence of the inclusion of a limited number of items. These two factor were also problematic in terms of item-to-total correlations, which were below the recommended 0.5 [28]. However, some authors advocate the cut-off value of 0.3 [35]. These factors were not removed from the model and the decision was based on content consideration and acceptable alpha values. The exclusion of items did not improve the model.

Another limitation is the limited response rate and the relatively high number of incomplete questionnaires. However, only two clinics out of 37 failed to provide at least one complete questionnaire. After the exclusion of 119 incomplete questionnaires, 131 cases were available for factor analysis of the selected 22 variables. As noted by Mundfrom et al. [36] there are no clear criteria for deciding the appropriate factor analysis sample size. Suggestions for minimum absolute sample size range from 100 to over 1000, and suggested minimum ratio of number of variables to number of factors range from 3 to 20, depending, among other things, upon the number of factors extracted and the level of communality [37]. The sample size in our study was limited, and we recommend further studies with a higher number of participants.

An EFA always produces a solution, but does not assess the risk that the EFA solution only describes the data set, and may not be generalizable. Due to these limitations, the results of our study may not be generalised to the whole population.

Based on the results in our study, we cannot state that the SAQ-AV is a reliable tool for measuring safety culture in the Slovenian out-of-hours care setting. This might be due to methodological limitations. Our study also showed that there might be other safety culture factors in OOHC not recognised before, such as psychological safety. We therefore recommend larger studies aiming to identify an alternative factor structure and other concepts of safety culture in OOHC.