Safety culture in the primary health care settings based on workers with a leadership role: the psychometric properties of the Slovenian-language version of the safety attitudes questionnaire – short form

Patient safety is an increasingly important part of quality assurance in health care. Within patient safety, a concept of a safety culture has developed which describes leader and staff interactions, attitudes, routines, awareness, and practices within an organisation [1]. As safety culture is a group, and mainly an organisational phenomenon, it differs across the different levels and organisations of health care [2, 3], so appropriate tools should be employed in order to capture the true state of affairs. Safety culture differs from safety climate the latter referring to the measurable components of safety culture, a kind of a snapshot of safety culture at the given moment in time [4].

Several tools have been developed to measure the safety attitudes of healthcare providers [5‐7]. Recent reviews of the Safety Climate Surveys in health care settings showed up to 17 [7, 8] available surveys, however only three were developed for a primary health care settings [7, 8]. Another systematic review on tools for safety culture in primary care [6] put forward the Safety Attitudes Questionnaire (SAQ) [9] as one of the widely used and also one of the most appropriate instruments for assessing patient safety culture. In 2007, the SAQ was adapted to outpatient (primary healthcare) settings [5]. This adapted version, the SAQ – Ambulatory Version (SAQ-AV), was found to be a reliable tool for comparing attitudes across different professional groups of healthcare providers outside hospitals [5, 10, 11]. Its shorter version, the SAO Short Form is being increasingly used in many countries [9, 12‐18].

In Slovenia, quality and safety at the primary health care level have been the focus of study over the past several years [19‐22]. Patient safety features in primary healthcare have been investigated through a study on the Quality and Costs of Primary Care in Europe (QUALICOPC), dealing with the organisation and accessibility of primary healthcare services [23]. In addition, safety culture has been measured by the Slovenian-language version of the SAQ-AV in out-of-hours primary health care settings [24‐26]. Out-of-hours settings in Slovenia differ from primary healthcare centres in terms of organisation, location, and staff. Out-of-hours health services are combined with emergency medical services and are available both at the same place and time. Sometimes, they are located in hospitals and sometimes in the healthcare centres. 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. An out-of-hours health centre has thus a unified leadership and can be seen as a “natural social unit”; which is a validation criteria for organizational climate measurements [27].

To date, the safety culture in primary health care settings such as primary health care centres has not been measured in Slovenia. Also, the SAQ – Short Form has not yet been validated in the Slovenian language.

Within this study we wanted to fulfil the following aims: 1) to test the reliability and validity of the Slovenian-language version of SAQ – Short Form in primary health care settings, and 2) to determine the factor structure of the Slovenian-language version of SAQ– Short Form in primary health care settings.

The Slovenian-language version of the SAQ – Short Form used with the employees of the largest primary health care centre in Slovenia did not fit perfectly to the original factor structure [9]. Our new factor model revealed six factors: 1) Perceptions of Management; 2) Stress recognition; 3) Teamwork Climate; 4) Communication; 5) Safety Climate; 6) Working Conditions and Satisfaction. Cronbach’s alphas of the factors were good, indicating no problems with the factors. The six-factor model of the Slovenian SAQ – Short Form covered 36 items.

Our factor structure differs from the original one [9] in the context of some factors. Factors Teamwork Climate, Perceptions of Management, Working Conditions, Safety Climate, and Stress Recognition are the same as in the original factor structure [9]. However, the items belonging to these factors differ (see Table 3). Our model also revealed one new factor (Communication) and one partly new factor (Working Conditions and Satisfaction). The latter is actually a combination of two factors from the original SAQ – Short Form scale. The factor Communication has not been revealed within the original scale but has already been revealed as an individual factor of the Slovenian-language version of the SAQ-AV [25]. Communication has also been recognised in other safety culture measurement tools as an important domain within the safety climate or culture [7, 10, 33].

Some items have negative loadings (all items in factors Teamwork Climate and Working Conditions and Satisfaction). This indicates that these factors are negatively correlated to the construct. It is a general agreement that the strength of a loading needs to be measured by its absolute value with the possibility that preconceptions about appropriate reverse coding might need modification based on the signs of factor loadings [34]. Therefore, we decided to keep the original wording of the items.

All the factors measured the same construct except for Stress Recognition. This could be due to the fact that other factors related more to work and the working environment, and Stress Recognition more to personality traits. Stress Recognition was weakly correlated with Communication, which indicates that if the working environment stimulates communication about safety issues, the workers feel that they manage stress better.

The validation studies on SAQ – Short Form in Norwegian ([35], Italian [13], Portuguese [14], Chinese [15], Swedish [16, 17], and German [18] language confirmed the original factor structure. However, they were carried out in the secondary health care level which makes it difficult to compare the results of our study to them as our study was done in primary health care settings. Our result of a CFA not perfectly fitting to the original factor model indicates that there are differences in the safety culture between primary and secondary health care levels which has already been pointed out elsewhere [6, 33]. Namely, primary health care differs from hospitals in terms of organisational structure, administrative and clinical processes and the reasons for encounters [36]. However, it could also indicate that the tool used (the SAQ – Short Form) was not specific enough for the primary health care settings and another tool could be used.

The safety climate and culture research in health care in general has been predominately conducted within acute care settings and hospital settings [7] therefore the existing measurements in primary health care settings are relatively new which makes our study an important one for this field of research.

The differences in factor structures between countries make cross-country comparisons of patient safety culture challenging. The structural differences may reflect cross-national variation in the nature and structure of primary care, or it may mean that item wordings trigger different connotations in different languages. For these reasons a wise option might be to compare countries at the item level rather than at the factor level [10].

The differences in the safety culture between the different levels of health care require different tools for measurements that would be developed for primary care or hospital safety culture measurement tools adapted for the use at the primary health care level. Since there is currently no consensus on which tool is the best to use [7], further research should focus on performing the \adaptations and validations of the existing tools for use at the primary health care level, with special emphasis on establishing the construct and criterion validity [7]. Another option would also be a development of a new scale which would be based on the theoretical model of the safety culture at the primary health care level.

Our study has some limitations that have to be mentioned. We are aware of the fact that an EFA always produces a solution, but does not assess the risk that the EFA solution only describes the data set, and it may not be possible to generalise. Due to this fact, the results of our study may not be generalised to the whole population. The sample size in our study was limited to employees with a leadership role, which could produce a selection bias and therefore limits the validity of our results. However, the results were very similar to our previous study [25] and also to the Dutch study [10] (both performed at the primary health care level) and this gives us confidence in the validity of our results. Although the response rate was high, there is no information on the characteristics of the non-respondents, so this could also be a source of bias.

We used the SAQ – Short Form for the primary health care settings despite the fact that there are other specific primary health care tools available [6, 7] and might have been a better tool for our study. This is another limitation of our study.

The results of our study suggests that the Slovenian-language version of the SAQ – Short Form with six factors could be a reliable and valid tool for measuring the safety culture in the primary health care workers with leadership role. The factors and item loadings in our study differed from the original SAQ – Short Form and the majority of other translated versions.

Further studies should explore the safety culture of other primary health care organisations, and perhaps also use other tools for measuring it in order to recognise other factors that are also important. Possibly, another tool for a more comprehensive measurement of safety culture that would be based in the theoretical frameworks should be developed.