The RNs working in the care of older people reported barriers related to the Setting and the Presentation subscales. The RNs proposed better availability of research reports in Swedish, additional time for research use and support from unit managers for enhancing research utilization. The research users among the RNs rated significantly less barriers on the three Nurse, Research, and Presentation subscales than the non-research users. In the following, we will discuss reported perceptions of barriers and the potential usefulness of the BARRIERS scale for identifying barriers to research utilization.
RNs' perceptions of barriers to and facilitators of research utilization
In general, the RNs working in the care of older people reported a rank order of barriers consistent with studies using the BARRIERS scale [
7]. However, more than 75% of the RNs in the present study rated five of the 30 potential barriers as actual barriers (Table
1). To compare with another Swedish study with RNs working at a university hospital, only two items were rated as actual barriers by more than 75% of the RNs [
21]. Thus, it appears as RNs working in the care of older people face more barriers than RNs working in hospitals. One reason might be that the average age of RNs working in the care of older people in Sweden is higher in comparison with RNs working in hospitals [
22], implying that a greater proportion of nurses working in the care of older people have an older nursing program and lack courses in research methods and nursing science. In the present study, the RNs with an older nursing program rated barriers on the Presentation and Nurse subscales significantly higher than the RNs having a more recent nursing program at the university level. These comparisons suggest that many of the RNs working in the care of older people do not have sufficient knowledge that facilitates research use.
Nearly all RNs working in the care of older people reported lack of knowledgeable colleagues and inadequate facilities for implementation as the major barriers to research utilization. The RNs suggested establishment of networks among colleagues, staff, researchers, and physicians for promoting research use. In previous studies, lack of knowledgeable colleagues is not common among the top ten barriers [
5,
7]. In the care of older people in Sweden, the settings are mostly small units (such as nursing homes and group dwellings) with the intention to be 'homelike' for the residents. The majority of nursing staff are ENs and NAs and in many smaller units there is only one RN employed. In Sweden, ENs and NAs have nursing training within the upper secondary school, which can be compared with RNs who since 1982 have a nursing program at university level. Recent national surveys have shown that nearly half of ENs and NAs working in the care of older people do not have adequate training [
22]. Moreover, smaller units have limited material and human resources for supporting practice development. These specific conditions in the care of older people are probably contributing to why so many RNs experience a shortage of knowledgeable colleagues. This finding is in accordance with a study among rural nurses who reported isolation and lack of a nursing research consultant as barriers to research utilization [
23]. These results point to specific barriers of a situational and geographical nature. The RNs working in specialist units in the present study perceived fewer barriers on the presentation and accessibility of research (the Presentation subscale) than RNs working in nursing homes. Other studies have shown that, in comparison with working in general settings, working in specialist settings enhance research use [
24].
The RNs' reports on lack of knowledgeable colleagues with who to discuss research may also be related to the English language. Two thirds of the RNs reported that the English language was a barrier to research use. In another Swedish study, almost half of the RNs working at a university hospital reported the English language and lack of knowledgeable colleagues as barriers [
21]. This 'second language' barrier has also been identified as a major barrier among RNs in other non-English-speaking countries (
e.g., Norway, Finland and Greece) [
25‐
27]. Generally, RNs prefer to use colleagues as information sources on research findings [
28,
29] and the need to have knowledgeable colleagues probably increases when the information is not in their own language.
Regarding the characteristics of the organization, many RNs reported lack of adequate facilities, lack of time to read and implement new ideas, and lack of support from other staff members as major barriers. These barriers have often been reported in previous BARRIERS studies [
5,
7]. Support from unit managers was one of the most frequently suggested factors to enhance the RNs' research use in the present study. Yet, respondents could not report lack of support from unit managers as a barrier because there is no item in the BARRIERS scale explicitly measuring the perception of support from front-line managers. One item is formulated 'Administration will not allow implementation', which, in the present study, was ranked as the 27
th barrier of 30. We interpret this finding to mean this item does not measure the concept of 'leadership'. Administration is a concept that in a Swedish context and language refers to and represents something impersonal and higher up in the organization. The wording 'not allow' also seems to be inappropriate to use in the care of older people. Such a setting often consists of small units and few staff categories, all of which implies a less formal organization. We consider it important to extend the BARRIERS scale with an item that measures support from front-line managers because the relationship between research use and leadership is well documented [
24,
30]. One positive finding was that in the present study the RNs did not report lack of authority to change practice as a major barrier. This finding is in contrast with what has been found in other studies [
5]. One of the advantages with working in smaller settings with less hierarchy might be that the RNs have more authority in influencing and putting evidence into practice.
Differences between reported research use and perceived barriers
A significant negative correlation was found between the RU index and the Presentation subscale, demonstrating that the RNs working in the care of older people that scored more use of research findings rated lower barriers related to the presentation and accessibility of research. These results differ from the findings of McCleary and Brown, who reported a significant correlation between research use and the Nurse subscale among nurses working in a pediatric teaching hospital [
9]. It is logical that if RNs do not have access to or do not understand the presentation of research findings in research reports, they will not use research findings in practice. Being aware of relevant research is the first stage in implementing findings according to Rogers' Innovation-Decision process [
8]. Access to research findings at the work place has also been identified as a determinant of research uptake in the care of older people [
31]. The statistical analyses demonstrated that the scale detected certain differences between research users and non-research users. The research users group rated significantly lower on three of the four subscales as compared with the non-research users group, indicating that the RNs using research perceived fewer barriers than those not using research (Table
3). These results support the underlying assumption of the BARRIERS scale,
i.e., lower perceptions of barriers imply more research use and visa versa [
6]. However, the lack of significant correlation between the RU index and the Setting subscale, and especially the lack of significant difference between the two groups' ratings on this subscale, is thought-provoking. The literature on barriers notoriously reports the Setting as the predominant barrier to research utilization [
5,
7]. Examining the two groups' ratings on the items in this subscale revealed that there was no consistent trend in the results (Table
4). The findings suggest that the Setting subscale measures heterogeneous characteristics of the organization, which appear to have different implications for research users compared with non-users. These differences are challenging to understand, especially when the goal is to identify adequate interventions for decreasing barriers to research use.
Is the BARRIERS scale useful to identify barriers to research utilization?
The BARRIERS scale measures perceptions of barriers regarding the nurse, the setting, the research, and the presentation of research. These four types of barriers to research use are, to some extent, congruent with the five types of barriers presented by Bosch and co-workers [
4], except from barriers related to patients and the wider environment (the structure), which are lacking in the BARRIERS scale. In the present study, the BARRIERS scale exposed differences in the perceptions of barriers to research utilization between research users and non-research users on the Nurse, the Research and the Presentation subscales, indicating that the instrument appears useful for identifying these types of barriers to research utilization. The lack of difference between the two groups on the predominantly reported Setting subscale undermines the validity of the BARRIERS scale to identify organizational barriers to research utilization. A further issue is that the instrument identifies barriers generally and with wide-ranging characteristics, making it difficult to design specific, tailored interventions to decrease the barriers. Previously, only one study has used the BARRIERS scale to identify barriers to actually promote research use in practice [
32]. Based on a pre-survey, Fink and co-workers implemented multiple organizational interventions, which ranged from integration of evidence-based practice philosophy into nursing job descriptions to establish unit-based journal clubs. The multi-faceted intervention significantly decreased the nurses' ratings on the Setting subscale, but which of the components of the organizational intervention that made this reducing effect was not possible to distinguish. This difficulty in designing specific interventions to reduce barriers is not unique. Findings reported by Shaw and colleagues [
3] and Bosch and co-workers [
4] point to a lack of useful theory for tailoring interventions to address barriers.
Methodological consideration
All RNs in participated municipalities were invited and a response rate of 67% was achieved, which must be judged as sufficient when using postal questionnaires [
33]. The study was performed in eight municipalities of varying sizes. These municipalities hold about one-third of the RNs working in the care of older people in Stockholm County. We believe our sample is representative for an urban region in Sweden. Conditions, such as turnover and lack of required training among staff, can differ between the care of older people in city regions and rural regions [
34] and our findings are probably not generalizable to all Swedish or international care of older people. The two questionnaires (BARRIERS scale and RUQ) have been used in several international and Swedish studies where they have been judged to be valid and reliable measures. In this study, the four subscales from the BARRIERS scale and the RU index from the RUQ were used. The validity of three of the BARRIERS subscales is supported by the current study. The reliability was tested using Cronbach's alpha statistics, and the measures were sufficiently consistent [
33]. The respondents' answers on the open-ended question on factors to facilitate research use supported their scoring on the barriers subscales,
i.e., they reported supportive factors mainly in accordance with their reported type of barriers. Some of the items in the research subscale in the BARRIERS scale had high proportions of no opinion responses. This finding has been reported in many studies using the BARRIERS scale [
5,
7] and indicates a lack of validity of this subscale.