Article Text

Effect of the Green Cross method on patient safety culture in a postanaesthesia care unit: a longitudinal quasi-experimental study
  1. Gørill Helen Birkeli1,2,
  2. Owen Matthew Truscott Thomas3,
  3. Ellen Catharina Tveter Deilkås3,4,
  4. Randi Ballangrud5,
  5. Anne Karin Lindahl1,2
  1. 1Division of Surgery, Akershus University Hospital, Nordbyhagen, Norway
  2. 2Institute of Health and Society, Department of Health Management and Health Economics, University of Oslo Faculty of Medicine, Oslo, Norway
  3. 3Health Services Research Unit, Akershus University Hospital, Nordbyhagen, Norway
  4. 4Department of Quality and Improvement and Patient Safety, Norwegian Directorate of Health, Oslo, Norway
  5. 5Faculty of Medicine and Health Sciences, Department of Health Sciences in Gjøvik, Norwegian University of Science and Technology, Gjøvik, Norway
  1. Correspondence to Gørill Helen Birkeli; gobi{at}ahus.no

Abstract

Background Hospitals should adopt multiple methods to monitor incidents for a comprehensive review of the types of incidents that occur. Contrary to traditional incident reporting systems, the Green Cross (GC) method is a simple visual method to recognise incidents based on teamwork and safety briefings. Its longitudinal effect on patient safety culture has not been previously assessed. This study aimed to explore whether the implementation of the GC method in a postanaesthesia care unit changed nurses’ perceptions of different factors associated with patient safety culture over 4 years.

Methods A longitudinal quasi-experimental pre–post intervention design with a comparison group was used. The intervention unit and the comparison group, which consisted of nurses, were recruited from the surgical department of a Norwegian university hospital. The intervention unit implemented the GC method in February 2019. Both groups responded to the staff survey before and then annually between 2019 and 2022 on the factors ‘work engagement’, ‘teamwork climate’ and ‘safety climate’. The data were analysed using logistic regression models.

Results Within the intervention unit, relative to the changes in the comparison group, the results indicated significant large positive changes in all factor scores in 2019, no changes in 2020, significant large positive changes in ‘work engagement’ and ‘safety climate’ scores in 2021 and a significant medium positive change in ‘work engagement’ in 2022. At baseline, the comparison group had a significantly lower score in ‘safety climate’ than the intervention unit, but no significant baseline differences were found between the groups regarding ‘work engagement’ and ‘teamwork climate’.

Conclusion The results suggest that the GC method had a positive effect on the nurses’ perception of factors associated with patient safety culture over a period of 4 years. The positive effect was completely sustained in ‘work engagement’ but was somewhat less persistent in ‘teamwork climate’ and ‘safety climate’.

  • Safety Management
  • Patient safety
  • Risk management
  • Organizational Culture
  • Postoperative Care

Data availability statement

Data are available on reasonable request. All the data generated are available for sharing on request.

http://creativecommons.org/licenses/by-nc/4.0/

This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.

Statistics from Altmetric.com

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

WHAT IS ALREADY KNOWN ON THIS TOPIC

  • Hospitals use multiple monitoring methods to get a comprehensive review of the types of incidents that occur. The Green Cross method is a suitable method for assessing the inherent safety of the organisation and may foster a restorative and just culture.

WHAT THIS STUDY ADDS

  • The Green Cross method had a positive effect on the nurses’ perception of work engagement teamwork, and safety climate over a period of 4 years. The effect on work engagement was completely sustained.

HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY

  • Hospitals stand to benefit from facilitating interprofessional reflexive spaces, such as in daily safety briefings. This may improve work engagement, learning and teamwork. More research to determine the effects of briefings on work engagement is called for.

Background

Globally, more than 10% of hospitalised patients experience adverse events.1 Some of the most common types of adverse events are related to surgery.1 Adverse events are unintended injuries caused by medical management rather than the underlying disease that prolong hospitalisation and cause disability or even death.1 This is an immense burden on individuals and consumes resources in health systems and societies that could be used productively elsewhere.2 About half of the adverse events can be prevented through a strategic and coordinated approach.1 3 One approach is to monitor incidents, including harmful events and near-misses, with the emphasis on making future care safer.3 4 It is imperative for healthcare professionals to get an overview of incidents to be able to prevent future harm. However, not all complications can be prevented; patients get older and frailer and react differently to medical treatment. By monitoring incidents, healthcare professionals can anticipate complications and respond more effectively when they occur, thereby contributing to resilience.3 5 6

To get a comprehensive review of the types of events that occur, complex adaptive systems should use multiple monitoring methods.3 7 The most widespread methods are incident reporting systems where healthcare professionals record incidents.4 8 However, incident reporting systems are not effective for double loop-learning, that is, cultural change.8 9 Under-reporting is common and linked to a lack of feedback and learning.10 Other barriers are non-consensus on what to report, complicated systems and fear of shame and blame.11 To make incident reporting systems more effective, it is advocated to decentralise their management from centralised hospital departments to clinical teams, make reporting easier, give timely feedback and focus more on learning.4 8 12–14 To promote learning it is imperative to develop a culture of respect, openness and transparency.3

The Green Cross (GC) method is a simple visual method for healthcare professionals to recognise incidents in real time.15 Contrary to traditional incident reporting systems, the GC method is based on teamwork where incidents are discussed in daily safety briefings and weekly quality improvement meetings, thus decentralising its management and providing daily feedback. The non-anonymous reporting requires psychological safety to dare to speak up about errors, thus facilitating openness and transparency.3 16 Moreover, it requires that everyone involved avoids blaming and shaming and supports a restorative and just culture.16 17 Templates and the use of incident definition facilitate clarity and promote learning.16 The GC method was invented in Sweden in 2011 and has since spread internationally.18 It consists of seven steps: (steps 1–3) daily interprofessional safety briefings in which incidents are collected, assessed and visualised, and (steps 4–7) the recording of adverse events in the hospital incident reporting systems, involvement of patient and relatives, and working on improvements.15

A culture of openness and learning is closely linked to the safety culture. The term was first used in 1986 to describe the cause of the nuclear power plant disaster in Chernobyl, which was attributed to a breakdown in the safety culture of the organisation.19 The aspects of safety culture that relate to patient safety are termed patient safety culture (PSC).20 PSC is the cornerstone of the safety movement21 22 and is perceived as an indirect measure of quality of care.23 24 A PSC can be defined as ‘an integrated pattern of individual and organisational behaviour, based on shared beliefs and values that continuously seeks to minimise patient harm, which may result from the processes of care delivery’ (EUNetPaS, p4).25 This definition implies active involvement in reducing patient harm, which is the crux of the GC method.

Positive PSC is statistically associated with reduced adverse events.26 It plays an essential role in the effective recognition and response to surgical complications and may be the main driver of the variation in failure to rescue.27–29 PSC as a part of safety culture can be measured and improved.30 Since PSC is a group-level property that emerges and lives in work units,31 32 initiatives to improve PSC at the unit level are necessary.27 Thus, the GC method was implemented in a Norwegian postanaesthesia care unit (PACU) in 2019, which is the context of this study.33 Questionnaires, such as the Safety Attitudes Questionnaire (SAQ),34 can effectively capture tangible themes of PSC, such as teamwork, error notification and learning, and staff well-being.35–37 Since 2018, Norwegian hospitals have used an annual hospital staff survey (staff survey) to assess the safety culture.38 This study explores the Staff Survey factors ‘work engagement’, ‘teamwork climate’ and ‘safety climate’. The factors were chosen because they play an important role in providing safe patient care.26

Related reporting methods like safety briefings have significantly improved PSC.39 The GC method has also significantly improved PSC in nursing units in a Swedish hospital.15 However, the longitudinal effects of these methods are not studied. Longitudinal studies are useful in evaluating the sustainability of interventions.40 This study fills the gap of longitudinal, controlled studies on the effect of the GC method and safety briefings on PSC,41 specifically in relation to teamwork, job satisfaction and work engagement.39 42 This may be useful to present and future users of the GC method and similar methods.

This study aimed to explore whether the implementation of the GC method in a PACU changed nurses’ perceptions of different factors associated with PSC over 4 years. The research question was How does the implementation of the GC method in a PACU change the nurses’ perceptions of ‘work engagement’, ‘teamwork climate’ and ‘safety climate’ over 4 years.

Conceptual framework

The conceptual framework for this study builds on the Systems Engineering Initiative for Patient Safety (SEIPS) model.43 SEIPS depicts how the GC method (input) shapes nurses’ work systems independently and collaboratively by providing a reporting tool that is simple to use and by attempting to improve PSC. In turn, work processes influence organisational, group-level outcomes (‘teamwork climate’, ‘safety climate’) and professional, individual-level outcomes (‘work engagement’). The inter-relatedness within and between each component illustrates the complexity of the system43 (see online supplemental file 1).

Supplemental material

Methods

Study design

The study used a longitudinal quasi-experimental pre–post intervention design with a non-equivalent comparison group.40 This design is useful when there are practical barriers to randomisation and when an entire unit receives the intervention and a similar unit is available that does not receive the intervention.40 44 The Transparent Reporting of Evaluations with Nonrandomized Designs (TREND) guidelines were used to ensure thorough and transparent reporting45 (see online supplemental appendix 1).

Supplemental material

Setting and sample

The intervention unit was a 25-bed general PACU in the surgical department of a Norwegian university hospital. During the study period, the intervention unit was staffed by 75–81 nurses, of whom roughly 50% were critical care nurses—that is, registered nurses with a clinical postgraduate education, with or without a master’s degree. The rest were registered nurses with a bachelor’s degree and at least 2 years prior clinical experience. The PACU cared for paediatric and geriatric patients, most of whom were postoperative. It also cared for adult intensive care patients as an overflow area. Patients usually stayed in the PACU for 2 hours postsurgery, sometimes up to 2–3 days. The nurse-to-patient ratio was 1:2. For critical care patients, the ratio was 1:1.

The units in the comparison group were selected based on similarities with the intervention unit.44 These included three nursing anaesthetists and surgical nursing units in the same surgical department that had not implemented the GC method. The units were staffed by 39–72 surgical nurses and nurse anaesthetists—that is, registered nurses with a clinical postgraduate education, with or without a master’s degree. The units cared for surgical patients who were handed over to the intervention unit postsurgery. The nurses usually cared for patients for a few hours during surgery. The nurse-to-patient ratio was mostly 1:1.

The units in both groups ran 24 hours a day, 7 days a week.

Intervention

A modified version of the seven-step GC method was implemented in February 2019, to better respond to context. Figure 1 shows how it was practised and some empirical experiences with the steps. The modification considered the time of recording the incident. This was done in real time by the nurse before the safety briefings (step 1). Modification also considered the daily safety briefings that were not attended interprofessionally, as this was not feasible. The recorded events were read aloud, and their severity was assessed and visualised using a template that included the days of a month assembled in the shape of a cross (step 3). At the weekly quality improvement meetings, the previous weeks’ events were discussed to develop a common understanding of patient risks and to learn from them (step 4). Although the managers prioritised the quality improvement meetings, it was not always possible to carry them out, especially on busy days.46 The meetings were attended by PACU nurses, certified nursing assistants and nurse managers. They were sometimes attended by one or two anaesthesiologists, pharmacists, surgical nurses and nurse anaesthetists, although their respective units did not use the GC method. More details regarding attendance are described elsewhere.46 Improvement work could be initiated based on these discussions and the monthly summary (step 5). The modification also regarded the patients’ and relatives’ involvement when patient harm occurred. This was not done because of the patients’ short stay in the PACU. Further details of the intervention are described elsewhere.33

Figure 1

The GC method as it was implemented, including empirical experiences with the steps. Original templates are reproduced with permission.15 GC, Green Cross; IRS, incident reporting system; QI, quality improvement.

Kotter’s eight-step process for leading change helped guide the implementation.47 In the preparation phase (1) the need for change was established, (2) a project group was created and (3) a modification of the GC method and a systems focus on patient harm were agreed on.48 In the action phase (4) the vision was conveyed to the nurses through 2-hour lectures, (5) all material was made available to empower the intervention, (6) small wins were celebrated and (7) the project group, including management, did not stop pressing on. To create a culture (8) daily safety briefings and weekly interprofessional improvement meetings continued. The managers even introduced the GC method to nurses during employment.46

Patient and public involvement

Patients and/or the public were not involved in the design, or conduct, or reporting, or dissemination plans of this research.

Data collection

PSC measurement

Reports from the staff survey were collected annually between 2018 and 2022 to explore the effects of the intervention.38 The staff survey was initiated by the national patient safety programme and regional health authorities. Its purpose is to improve the work environment and patient safety in the units, departments and the whole organisation.

The staff survey combines statements (items) regarding health, safety and environment with validated items from the Norwegian-translated SAQ34 and the general Nordic Questionnaire for Psychological Social Factors at Work.49 This includes 9 factors made out of items with similar meanings, comprising 41 items.38 Six factors were beyond the scope of this study and were, therefore, excluded. For details regarding exclusion, see online supplemental file 2. Three factors derived from SAQ were included, comprising 17 items with 5–6 items per factor (table 1).

Supplemental material

Table 1

Included factors and single items from the staff survey

All items were scored on a 5-point Likert scale (1=strongly disagree, 2=disagree, 3=neutral, 4=agree and 5=strongly agree). A higher score represents a more desirable result. The mean score for each item was converted from a 1–5 scale to a 0–100 scale.38 The summarised values were then divided by the number of respondents to obtain the aggregated data at the mean unit level.

The analysed results were reported back to the hospital units if they had more than five respondents, to protect their anonymity.38 The reports included the following: (1) the number of sent invitations, the number of answers and the answers in per cent, (2) the mean score (0–100) for each factor and (3) the percentage of staff with a positive response (ie, ‘I agree’ or ‘I strongly agree’) (≥75 scores).

Procedure

The staff survey was distributed to all staff with an employment relationship of more than 3 months.38 It was distributed in March every year, with 4 weeks to respond. A link to the software survey platform (euro.confirmit) was distributed via work email or a text message to each employee. The survey was answered individually during work time or at home, and it took less than 10 min to complete. The nurses were encouraged to answer the survey by their respective managers. Reminders were sent by mail and text messages during the survey period. The responses were identified with only unit affiliation and were thus anonymous.

Statistical analysis

All analyses were performed using R statistical software (V.4.3.2).50 A logistic regression model was used to analyse the relationship between the intervention and the relevant outcomes from the Staff Survey.51 The response variable used was the binary measure of whether each individual responded with a positive answer (either ‘agree’ or ‘strongly agree’) to each question on the Likert scale. This threshold was chosen because it was the data summary accessible from the available resources.

Data were used for all years available in both the intervention unit and the comparison group—that is, 2018–2022 inclusively. The model was constructed with a factor structure for the 4 years (using 2018 as a reference category), a variable representing unit membership (using the comparison group as a reference category), and an interaction effect between unit membership and time.

The effects reported were the regression coefficients of the logistic regression model alongside standard errors and p values. P values of ≤0.05 were considered significant. The effect sizes can be interpreted as log odds ratios (ie, an effect size of zero is no effect), and their absolute values were interpreted as follows: 0 <‘small’ < 0.4 < ‘medium’ < 0.7 < ‘large’ < 1.4 < ‘very large’. The OR was the relative probability of providing a positive answer versus a negative or neutral answer in one group divided by the relative probability of providing a positive answer versus a negative or neutral answer in the other group. The effects reported as ‘(intercept)’ in table 3 represent the log odds in the comparison group in 2018. The effects reported as just years (eg, ‘2019’) can be interpreted as being the variation in log OR with time present in the comparison group relative to 2018. The effects reported as just ‘units’ can be interpreted as the difference between the groups at baseline in 2018, with the comparison group as a reference.

The effects reported as an interaction between units and time (eg, ‘units:2019’) describe the variation in the intervention unit, in addition to the difference at baseline with the comparison group and the variation in time in the comparison group. Put differently, it is the difference in the log OR between the groups at each time point relative to 2018, correcting for baseline variation.

Considering that the intervention happened at the start of 2019, the interaction effects with units and time can be interpreted as the result of the intervention, under some assumptions, and therefore, the main results of the analysis.

Results

Descriptive results

In the period of 2018–2022, a total of 277 individual survey responses were completed in the intervention unit (mean 55 surveys /year), and 565 individual survey responses were completed in the comparison group (mean 113 surveys /year). The average response rate in the intervention unit ranged from 49% to 77% (mean 71%). The lowest response rate, in 2020, coincided with the beginning of the first wave of the COVID-19 pandemic. The response rate in the comparison group ranged from 67% to 84% (mean 75%).

The descriptive statistics of ≥75 scores of all the included factors in all years are presented in table 2 and figure 2. This score refers to the distribution within the unit and to an extent to the degree of consensus, which has been found to correlate with patient safety.52

Table 2

Percentage of respondents who answer positively (≥75 scores) on each factor

Figure 2

Work engagement, teamwork climate and safety climate. Proportion of respondents in the intervention unit (IU) versus comparison group (Comp) that answers positively (≥75) on each factor. Vertical line marks the start of the Green Cross method implementation (February 2019).

Logistic regression results

For the ‘work engagement’ score, there was no substantial variation in the comparison group with time or between the two groups at baseline (table 3). In the intervention unit, there was a medium-to-large positive influence of the intervention in 2019, 2021 and 2022, with statistically significant p values.

Table 3

Logistic regression results

For the ‘teamwork climate’ score, there were substantial increases in the comparison group with time, with significant and medium-to-large increases in 2020, 2021 and 2022, relative to 2018. There was no difference between the groups at baseline in 2018. In the intervention unit, there was a large positive significant effect of the intervention in 2019 relative to 2018 but not in subsequent years.

For the ‘safety climate’ score, there was a small negative effect in the comparison group relative to the baseline in 2019 and a medium-sized positive difference between the groups at baseline. In the intervention unit, there were large increases in the outcome score relative to the baseline in 2019 and 2021, in addition to the other observed variations.

Discussion

As related to the SEIPS model, the improved outcomes with time in the intervention group, with a medium-to-large effect size, indicate the effect of the GC method.53 The results suggest that the greatest change in PSC was achieved in the first month after the implementation (2019). In 2020, during the first wave of the COVID-19 pandemic, all factors returned to the baseline level. Due to the low response rate, this result might not have been representative of the culture.54 Nevertheless, the GC method was used to a lesser degree during this period because of the strains of the pandemic and could explain the lower score.46 The decrease, followed by the significant large increases in ‘work engagement’ and ‘safety climate’ in 2021, indicates its effect on PSC. After 4 years, ‘work engagement’ remained significantly increased, suggesting that the effect was only partially sustained and that the intervention perhaps affected the individual level more than the group level.

There was a significant, sustained, medium-to-large increase in ‘work engagement’, which strongly suggests that the GC method increased work engagement. This can be regarded as an improved professional outcome in SEIPS. Healthcare professionals’ well-being is equally as important as performance in SEIPS.55 In this study, ‘work engagement’ includes job satisfaction, having engaging work tasks, being able to develop oneself through work and receiving constructive feedback (table 1). The findings matched the nurses’ commitment to patient safety and to improving the system, which was sustained 3 years after the implementation of the GC method.46 The results also agree with previous studies in which safety briefings were found to increase job satisfaction.42 A possible explanation for the increased engagement may be the novel safety briefings and weekly meetings, which allowed the nurses to participate in decision-making and communicate openly with their managers. While most PACU nurses attended the daily safety briefings on a regular basis, they attended the weekly quality improvement meetings less frequently due to shift work and an unpredictably busy unit.46 This suggests that the safety briefings influenced the nurses more than the improvement meetings. Job satisfaction is strongly linked to psychological safety, which is a prerequisite for being willing to contribute ideas and actions and is thus essential for the GC method.56 Discussing patient safety questions and suggesting improvements might have increased the nurses’ autonomy, contribution and sense of belonging. These are important steps to promote well-being and reduce burnout.57 Involving the team in decision-making and creating a psychologically safe climate is also associated with a positive PSC.35 Healthcare professionals’ well-being affects patient experience.58 Therefore, increasing ‘work engagement’ should be the prime consideration for every manager, and the GC method could help facilitate this.

‘Teamwork climate’ can be regarded as an improved organisational outcome in SEIPS. The initial significant large increase in the ‘teamwork climate’ score may be explained by this being the first time patient safety questions could be discussed by healthcare professionals across ‘silos’—that is, the operating room and the PACU. The interprofessional team experienced that the different perspectives conveyed in these meetings improved interprofessional communication and increased the shared understanding of the risks and safety in the perioperative pathway.46 The finding also agrees with previous studies in which safety briefings and the GC method had a positive effect on teamwork.15 42 However, in the present study, the large positive significant effect on ‘teamwork climate’ score was not sustained. A plausible explanation is that interprofessional attendance was achieved only to a limited extent.46 This was partly due to organisational factors, such as a lack of time.46 Interestingly, the comparison group had a substantial increase in ‘teamwork climate’ with time. A possible reason may be that a contamination effect occurred and that the scores were interrelated.44 Creating reflective spaces across an organisation is the foundation of resilient capacity in healthcare.59 Interprofessional reflexive space is key for learning, developing a collaborative culture and mutual understanding.59–61 It strengthens relationships and teams, enhances PSC and improves patient care.39 Moreover, good collaboration can positively influence patients and is vital for good recovery and successful surgery.62–64 Therefore, hospitals stand to benefit from facilitating interprofessional reflexive spaces, such as briefings, for the perioperative team.

‘Safety climate’ may be considered an improved organisational outcome in SEIPS, and this score increased significantly in 2019 and 2021, with a large effect size. Safety climate includes reporting and learning from errors (table 1). The nurses reported more near-misses and had increased communication and learning about near-misses after implementing the GC method.33 46 This concurs with previous studies.7 15 A culture of trust and dialogue, in which it is safe to talk about mistakes and there is collective learning, is a construct associated with a positive PSC.35 This is also the hallmark of the GC method.16 Nevertheless, in 2022, there was no significant increase in ‘safety climate’. The nurses were reluctant to report because of limited visible improvements, for example.46 Insufficiencies in reporting and planning improvement work were also found in other Norwegian hospitals that use the GC method.65 Learning is a complex social process that includes actively reflecting on and reorganising shared knowledge, organisational routines and practices.13 66 67 Interestingly, in Sweden and Norway, injuries that occurred in other units were problematic because they could be considered as looking for mistakes in others.16 33 46 This could have made it difficult to work on system improvement. To increase learning, it is important to collect data in real time and discuss this within the interprofessional team in the closest proximity to the point of care.59 68 However, investigations and improvements targeted at the unit level may give a limited perspective on system failures and provide fragmented and limited quality improvements.69 To make viable improvements, clinicians must be trained in human factors and systems engineering.70 A secondary support structure, such as hospital boards, may be beneficial in assisting in the trying out of new quality improvement ideas.71

In Norway, postoperative death following a surgical complication has dropped around 30% the last decade, without a similar fall in surgical complications.29 This suggests that the safety culture in most Norwegian PACUs has improved. Nevertheless, there are significant differences in postoperative deaths between hospitals, although they have similar complication rates.29 72 Differences in PSC may explain this variance.27 29 In this study, the mean percentage of nurses who responded positively to all three factors increased from 77% to 89% over 4 years (table 2). A positive consensus of above 60% on the safety and teamwork climate is considered a mature PSC in Norway,73 based on recommendations from 15 years ago.74 Perhaps it is time to raise the bar and aim as high as 80% to further reduce postoperative mortality.

Strengths and limitations

The study has several strengths: The longitudinal design with pretest and multiple post-test measurements, and the use of a comparison group.75 It took place in one setting over 4 years, allowing for an in-depth investigation of processes in one place. The richness of data captured in one setting allowed a detailed examination of the phenomenon across 4 years. The use of several post-test measurements provided evidence to refute the regression to the mean as an alternative explanation.75 The results appear to be consistent with previous qualitative findings in the intervention unit. The staff survey could be considered a substantial source of information because it is used as a national quality indicator of PSC.76

This study also has some limitations: First, there was a significant difference between the groups regarding ‘safety climate’ score at baseline. The lack of randomisation (ie, randomly allocating units to intervention or comparison groups and then treating them identically apart from the intervention), threatens the internal validity and makes it challenging to make inferences about causality.75 The results could have been affected by contamination, by secular trends that occurred at the same time, or by other elements in a complex system. PSC could have matured over time, and this could have been confused with the effect of the intervention. Second, the survey itself was not validated. Third, the first author previously helped implement the GC and had interviewed the nurses. This in-depth knowledge could be considered both a strength and a limitation of this study. Fourth, as a modified version of the GC method was used, the results could have been different from those of the original version. Fifth, other outcomes of the SEIPS model, such as adverse events or the number of events reported in the hospital’s incident reporting system, were beyond the scope of this study and therefore not included. This may be worth investigating in future intervention studies. Lastly, the research took place during COVID-19 which undoubtedly also affected the study procedures and findings. Therefore, caution must be taken in generalising the results.

Conclusion

The results suggest that the GC method had a positive effect on nurses’ perceptions of areas of PSC over a period of 4 years and could thus be used to achieve a positive PSC in a postoperative context. However, the effect was only partially sustained. The decline in ‘teamwork’ and ‘safety climate’ could indicate insufficient utilisation of the GC method regarding interprofessional safety briefings and work on improvement. It could also be caused by other elements in a complex system. The research took place during COVID-19 which undoubtedly affected the study procedures and findings. Therefore, future studies focusing on how to improve interprofessional learning and improvement work in the perioperative pathway are desirable. We also recommend further research to determine the effects of briefings on work engagement.

Data availability statement

Data are available on reasonable request. All the data generated are available for sharing on request.

Ethics statements

Patient consent for publication

Ethics approval

The hospital data protection officer approved the study (Ref. 2022_37). Privacy regulations do not apply because the information is anonymous and thus cannot be linked to a specific person. Approval was obtained from the hospital, and the units involved were informed. Completion of the survey was regarded as informed consent.

Acknowledgments

We would like to thank the units involved and the university hospital for letting us use the Staff Survey for this study. We thank the two referees for valuable comments, which have made this a stronger paper.

References

Supplementary materials

Footnotes

  • Contributors GHB conceptualised and led the study, conducted the data collection, contributed to the data analysis and interpretation and wrote the article. OMTT performed the formal data analysis and contributed to writing the ‘Statistical analysis’ and ‘Logistic regression results’ subsections of the paper. ECTD, RB and AKL conceptualised the study and contributed to data analysis and interpretation. AKL supervised and contributed to all parts of the study. All authors critically reviewed the original draft and the final version and approved the final version to be published. GHB is the guarantor of this contributorship statement.

  • Funding The Division of Surgery, Akershus University Hospital, Norway, financed this study.

  • Competing interests None declared.

  • Patient and public involvement Patients and/or the public were not involved in the design, or conduct, or reporting, or dissemination plans of this research.

  • Provenance and peer review Not commissioned; externally peer reviewed.

  • Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.