We used a cross-sectional design based on the Norwegian version of the AHRQ’s NHSOPSC to assess staff perceptions of patient safety culture in a sample of nursing homes [17] [21] . Twelve Norwegian nursing homes from six different municipalities in southern and western Norway representing both urban and rural districts were purposively selected using variation as the main selection criteria. Variation was related to the size of the nursing homes, ranging from one to five wards. Staff participants, according to inclusion criteria, ranged from 15 - 172 per nursing home. Additionally, the nursing homes differed in how they were organized depending on size, but also related to their level of integration with daycare and home based services. The sample included nursing homes offering specialized care such as long- and short-term care, subacute care, rehabilitation, care for patients with cognitive impairments (such as dementia) and palliative care. Based on the variation criteria the sample of nursing homes included would to a certain extent represent the variety of nursing home settings in Norway. Information regarding the study and the survey instrument were presented to nursing home managers as part of the recruitment process. The sample consisted of healthcare staff who were employed in a minimum of 30% of positions and spoke the Norwegian language. A total of 671 paper-based questionnaires were distributed to study participants between June and September 2013. An information letter followed the questionnaire and participation was based on written informed consent. Each nursing home had a contact person who was responsible for local administration of questionnaires.

Healthcare in Norway is a public responsibility that is mainly financed through taxation and public sources. While the state is responsible for specialized healthcare, such as hospitals, the municipalities are responsible for primary care, including nursing homes. Nursing homes play an important role in the health services system by offering advanced care, including long- and short-term care, sub-acute and acute care, rehabilitation, care for patients with dementia and cognitive impairments, and palliative care [22] .

The Norwegian Coordination Reform, which was implemented in 2012, aimed to reduce the need for costly specialized healthcare by transferring medical treatment and nursing tasks from hospitals to nursing homes [23] . Consequently, nursing homes have received an increased number of patients with poor functioning and who are at risk for adverse events. Through the national patient safety program, leadership and safety culture are regarded as crucial aspects for improving patients’ safety [24] .

Safety culture was measured with the Norwegian version of the NHSOPSC inventory, which consists of 43 items covering the perceptions of healthcare staff regarding patient safety culture. The original instrument, which comprises 12 factors, has been tested and validated on a large scale in the United States, and is used to examine patient safety culture ratings [17] [19] . When the NHSOPSC inventory was translated and tested in a Norwegian context, results from confirmatory factor analysis showed that a 10-factor model had an acceptable fit in the Norwegian setting as follows: root mean square error of approximation = 0.060, comparative fit index = 0.934, Tucker-Lewis index = 0.926, χ²  =  2058.33, degrees of freedom = 765, p < 0.001, and acceptable factor loadings from 0.40 - 0.89 [21] .

All NHSOPSC items are rated on Likert scales from 1 - 5, and include a response alternative “does not apply (DA) or “don’t know” (DK). In addition to the 10 dimensions (Table 1), we included one overall rating question considered as outcome as follows: “Please give this nursing home an overall rating on patient safety” (scale from 1 - 5). The survey also collected background demographic variables, such as staff position or background, number of years in a nursing home, work hours per week, work shift, and working directly with patients most of the time [17] [21] .

The study was approved by the Norwegian Social Science Data Services (Ref. No. 2012/32450) and the Norwegian Regional Committees for Medical and Health Research Ethics (Ref. No. 2011/1978). Participation was based on written informed consent.

We used SPSS Version 23.0 (IBM Corporation, Armonk NY, USA) for Windows for the analyses. Average percent positive responses for each of the 10 safety culture dimensions were calculated by averaging the item-level percentage positive responses, when the response alternative DA or DKwas excluded [17] . When the responders answered “strongly agree/agree” and “most of the time/ always” or “strongly disagree/disagree” or “never/rarely” for reverse worded items, the answers were categorized as a positive response. Items with the corresponding mean, standard deviation (SD), number of positive responses per item, percentage of positive responses per item and, average percentage per dimension were reported.

Differences in perception of patient safety culture (10 dimensions) according to educational background and position (managers included leaders at the first-line level, healthcare workers with a minimum of a Bachelor’s degree and healthcare workers with upper secondary school education and assistants) were tested using one-way, between groups analyses of variance (ANOVA). Assistants included those who responded “assistants” and those who responded “others” related to staff position or background. Separate post-hoc tests (LSD) were conducted for parameters that were significant in the overall analyses.

Multiple linear regression analyses were performed to examine the predictive value of the 10 safety culture dimensions on the outcome question “Please give this nursing home an overall rating on patient safety”. All of the independent variables were entered simultaneously. Pearson’s Product―Moment Correlation and preliminary analyses were conducted to ensure no violation of the assumptions of normality, linearity, multicollinearity, and homoscedasticity before we performed the regression analyses. No violation was found and the correlation coefficients varied between 0.236 and 0.635. The unstandardized coefficient (B), standard error of the mean (SE), standardized beta coefficient (b), and the unique predictive contribution for each of the dimensions (sr²_unique) are presented. A p value of <0.005 was regarded as significant.

A total of 466 healthcare staff responded to the questionnaire, with a response rate of 69%. The participants were representative for Norwegian nursing home staff and 6.3% included managers, comprising leaders at first line level (n = 29), 39.6% were healthcare workers with a minimum of a Bachelor’s degree (n = 181), and 54% were healthcare workers with upper secondary school education and assistants, (n = 247). Furthermore, 45.7% of the staff reported working >10 years in the nursing home (n = 209), 71.8% reported work hours >25 per week (n = 326), 67.7% worked in the daytime most often (n = 303), and 95.2% of the respondents working directly with patients most of the time (n = 436).

The proportion of positive responses was high. Six of the 10 dimensions had an average percentage of positive responses above 70% (Table 2). The dimensions with the highest percentages were “supervisor expectations and actions promoting patient safety” (88%) and “feedback and communication about incidents” (87%). “Non-punitive response to mistakes” also showed a high score (78%). The dimensions with the lowest percentages were “staffing” (46%) and “training and skills” (56%). Moreover, 80.9% of respondents rated the overall patient safety grade in their nursing home as excellent or very good, 17.4% as acceptable, and 1.8% as fair (n = 449).

Managers, including leaders at the first-line level, reported higher scores in all of the dimensions, except for “compliance with procedures” compared with the

Notes: R = reverse coded items. Positive responses were defined as answering “strongly agree/agree” and “most of the time/always” or “strongly disagree/disagree or “never/rarely” for reversed coded items. ¹Average percentage of positive responses per dimension. A, B, C and D representing sections in the questionnaire.

two other staff groups (Table 3). With regard to educational level, healthcare workers with upper secondary school and assistants reported higher scores on two dimensions, “handoffs” and “management and organizational learning”, compared with healthcare workers with a minimum of a Bachelor’s degree.

Standard multiple linear regression analysis showed that the 10 NHSOPSC dimensions explained 47.2% of the variance for the outcome question “Please give this nursing home an overall rating on patient safety” (F [10, 384] = 34.39, p < 0.001) (Table 4). Only three dimensions (“staffing”, “handoffs,” and “management and organizational learning”) made unique significant contributions to the model when the overlapping effects of all other variables were removed. “Management and organizational learning” made the strongest unique contribution to the model and explained 28.1% of the variance for the outcome question “Please

give this nursing home an overall rating on patient safety”.

Differences between groups were tested with ANOVA and post-hoc analyses were performed with LSD. *p < 0.05; **p < 0.01. ^aComparison between managers and healthcare workers with a minimum of a Bachelor’s degree. ^bComparison between managers and healthcare workers with upper secondary school education and assistants. ^cComparison between healthcare workers with a minimum of a Bachelor’s degree and healthcare workers with upper secondary school education and assistants.

*p < 0.05; **p < 0.01. B = The unstandardized coefficient, SE = standard error of the mean b = Standardized beta coefficient, sr²_unique = the unique predictive contribution for each of the dimensions.

This study is the first using the NHSOPSC in a cross-sectional design to report how a sample of healthcare staff in 12 Norwegian nursing homes perceive patient safety culture. Our study showed a positive awareness of patient safety culture among nursing home staff. Frequently cited dimensions related to safety culture, such as “leadership commitment to safety”, “non-punitive response to mistakes”, “teamwork”, and “communication openness”, had a high percentage of positive responses [13] . To some extent, these findings contradict previous studies indicating that patient safety culture is poorly developed in nursing homes and LTC settings, including a recent Norwegian study [1] [2] [3] [6] .

Our study supports Castle et al.’s study that compared safety culture in nursing homes and hospitals in the US using respectively the NHSOPSC and the HSOPSC [25] . They found that nursing home scores were generally better when comparing similar items in the HSOPSC questionnaire. Out of 26 highly similar items in the questionnaires, 16 of the NHSOPSC scores were higher than the HSOPSC scores. These results from US were not confirmed in a study from Netherland where patient safety culture in nursing homes, measured with SAQ, were comparable to those found in Dutch and US Intensive Care Units and ambulatory services [7] . Due to the scarcity of studies from European nursing home context using the NHSOPSC, cautions should be made when comparing studies having used different assessment tools.

In our study, the dimensions “supervisor expectations and actions promoting patient safety” (88%) and “feedback and communication about incidents” (87%) showed the highest percentages of positive responses. These results indicate that staff working at the bedside in Norwegian nursing homes have confidence in their frontline nursing managers. Engaged nursing managers listening to frontline staff’s concerns and promoting improvement of patient safety can serve as important role models in nursing homes. A recent Danish study investigated staff’s perceptions of patient safety culture before and after a leadership intervention [26] . This Danish study showed that strengthening leadership could act as a catalyst for improvement in the proportion of staff with positive attitudes and a more positive safety culture.

Fear of reporting has been recognized as a barrier to improvement in safety nursing homes [10] . Data from the AHRQ Comparative Database Report 2016 still found “non-punitive response to mistakes” among the dimensions that had the lowest average percentage of positive responses (54%) [27] [28] . In contrast, our study showed that “non-punitive response to mistakes” had a high average percentage of positive responses (78%). This finding indicates that staff feel confident in reporting safety concerns. Furthermore, teamwork, identified in the literature as the second most critical sub-dimension of patient safety culture, after leadership, showed a high average percentage of positive responses in our study (76%) [12] . Norwegian work life is characterized by a democratic leadership style and staff involvement in decision-making [29] . Additionally, national patient safety initiatives and legislation have shifted from focusing on the person to focusing on the system [24] . Notably, we found that “communication openness” had a slightly lower average percentage of positive responses (71%) than “non- punitive response to mistakes” (78%) and “teamwork” (76%). This is in line with a Swedish study that described nurses’ views on safety in nursing homes, and identified lack of communication and negative attitudes to reporting incidents among profound patient safety barriers [8] .

Our results for the dimensions “staffing” and “training and skills” are of concern because they had the lowest percentages of positive responses. Simultaneously, there has been steady growth of older patients with multiple diagnoses who are discharged from hospitals to primary care services, including nursing homes [30] . A recent study among municipality healthcare nurses in Norway concluded that the majority of nurses experienced complex and challenging work, but they lacked time, competencies, equipment, and information to sufficiently care for their patients [30] . Staffing appears to be a common challenge in nursing homes, exacerbated by financial issues and a lack of skilled staff and training [2] [30] [31] . A study that compared nursing home staffing standards in six countries showed wide variations within and across countries [16] . The increasing levels of severity of illness among patients in nursing homes highlights the need for further attention to staffing standards, to assure the quality of care. Norway has no official standards for staffing in nursing homes [31] . Staff training is an important aspect of safety improvement, but may be challenged by work shift arrangements and insufficient staffing levels.

The respondents in the current study were considered representative for nursing home staff and nearly all of them (95.2%) worked directly with patients most of the time. In our study sample, we found that managers, including leaders at the first-line level, reported higher scores on all the dimensions, except for “compliance with procedures”, compared with the two other staff groups (healthcare workers with a minimum of a Bachelor’s degree and healthcare workers with upper secondary school education and assistants). This finding is in line with previous studies, which reported that managers have significantly more positive safety culture perceptions than do frontline staff [3] [19] [20] . Comparison with other European studies is not possible as they lack management position in background characteristics” [6] [7] .

Educational level appeared to have less influence on staff’s perceptions of patient safety culture than management position. Still, we found that healthcare workers with upper secondary school education and assistants reported higher scores in the two dimensions “handoffs” and “management and organizational learning” than did healthcare workers with a minimum of a Bachelor’s degree. These results are in line with a Dutch study using SAQ reporting a negative correlation between education and safety climate [7] . The SAQ factor safety climate seems to some extend have similarity to the management and organizational learning dimension in the Norwegian NHSOPSC. Lower scores among staff with a minimum of a Bachelor’s degree might be a result of higher clinical competence and clinical assessment ability. However, the results to some extend contradictory to the findings of a recent Norwegian study where neither profession nor work experience were significantly associated with mean scores for any patient safety factor [6] . This inconsistency in results might be explained by study samples and should be further investigated.

The present study showed that the “management and organizational learning” dimension was the strongest unique contributor to the outcome question related to overall staff ratings on patient safety. This dimension consists of 10 items, including management support, safety priority, learning systems, and improvement changes. Considering that organizational characteristics are a stronger predictor of safety culture than individual behavior, changes toward an improved patient safety culture in nursing homes should be performed at an organizational level [10] . Management support means putting patient safety issues on the strategic agenda. Staff in nursing homes should know that they are not held accountable for system failures. Learning systems should be able to capture information and concerns from the frontline staff, besides having the capacity to drive improvement. The patient safety factor “perceptions of management” including managements support of daily efforts seems not to be included in the Norwegian version of SAQ for nursing homes [6] . Comparison is therefore not possible. Hence, previous studies have suggested a need for leadership involvement at all organizational levels and highlight managers’ responsibility in establishing and sustaining safety culture as a key driver in patient safety work [24] [26] [32] .

A strength of the present study is the response rate of 69%, which can be partly explained by the substantial leadership involvement in recruitment and feedback of results. The high number of respondents working directly with patients (including day, afternoon, and night shifts) provided important information from frontline staff on patient safety culture in nursing homes. Respondents included staff with relatively high levels of education and low rates of turnover, which are considered important to patient safety.

Twelve nursing homes located in the southwest region of Norway were included in the study based on their interest in patient safety issues and by actively enrolling in the study. Although our study results are not necessarily generalizable to all nursing homes in Norway, they provide a valuable overview of the patient safety culture across a number of nursing homes. Some of the nursing homes in our sample also participated in the Norwegian Patient Safety Campaign, adding explanatory value to the high percentages of positive responses.

One limitation of this study is the low number of managers participating in the study (n = 29). Another important limitation is the lack of an objective outcome that measures patient safety (e.g., falls and medication errors). Therefore, we had to relate to the subjective overall rating question “Please give this nursing home an overall rating on patient safety”.

Assessing patient safety culture can be the first step in identifying areas of improvement and barriers to provide safe care. We found a positive awareness of patient safety culture among frontline staff in participating Norwegian nursing homes. The highest proportion of positive responses was for the dimension “supervisor expectations and actions promoting patient safety,” indicating that staff have confidence in their frontline nursing managers. Furthermore, “feedback and communications about incidents” and “non-punitive response to mistakes” had high scores. Fear of reporting mistakes is considered a major barrier to improvement in safety. Areas of particular concern were staff perceptions of the “staffing” and “training and skills” dimensions, especially considering the increasing transfer of medical treatment and complex nursing tasks from hospitals to nursing homes.

As expected, managers reported higher scores on all dimensions, except for “compliance with procedures”, compared with other staff groups, while educational level had less effect on staff’s perceptions of patient safety culture. “Management and organizational learning” made the strongest unique contribution to overall staff ratings of patients’ safety, indicating the importance of actions taken at an organizational level to support a robust safety culture. Further research should include objective outcomes that measure patient safety, such as falls and medication errors, to identify if improvement in safety culture affects clinical outcomes. Culture assessments may increase the awareness of safety issues and support quality improvement initiatives.