A Resilient Approach to Patient Safety called Safety II: A Systematic Literature Review

A Resilient Approach to Patient Safety called

Safety II: A Systematic Literature Review

Hilde Jalink S3780708

MSc Business Administration: Health Faculty of Economics and Business

University of Groningen

Supervisor and Co-assessor University Dr. S. Täuber & Dr. J. de Bloom

Supervisor UMCG, Groningen Prof. dr. J.E. Tulleken

21st _{of June 2020, Groningen}

2

Abbreviations

CRM Crew Resource Management

DISCERN Debriefing In Situ Conversation after Emergent Resuscitation Now FRAM Functional Resonance Analysis Method

NDM Naturalistic Decision Making RAG Resilience Analysis Grid RE Resilience Engineering

RETIPS Resilience Engineering Tool to Improve Patient Safety RMF Resilience Mapping Framework

WAD Work-as-Done

3

Abstract

Background: The main objective of this study is to contribute an overview of the existing literature on the implementation of Safety II. This new view on patient safety is originated from the concept of Resilience Engineering, which assumes that people can adjust their performance to their environment and its conditions.

Method: A systematic literature review was conducted to find relevant research papers. Three search engines, ‘EBSCO’, ‘PubMed’, and ‘Web of Sciences’, were used to find relevant articles. The articles had to meet a number of selection criteria in order to be included in the review: 1) Based on studies with primary and secondary data, 2) written in English, 3) be available online, and 4) focus on resilience in healthcare, not other non-related health sectors. The articles also had to pass a quality assessment. Content analysis is used to extract data from the articles.

Results: Twenty general findings were found in 22 included articles. The findings were classified into three categories: A) resilience-oriented models, B) reason for implementation, and C) facilitators and barriers to implementation.

Conclusions: Resilience can be implemented using several methods or models. Organizations and healthcare providers have different reasons to strengthen resilience performance. Seven main facilitators have been identified, which can also be experienced as barriers if not handled properly. Future research should show how resilience and Safety II can be implemented in people’s performance who do not work at the front line and it could be further explored in other sectors than healthcare.

4

Introduction

At the moment of writing, the world is in the grip of a pandemic. The coronavirus broke out in December 2019 in China, and within 2.5 months, it spread to more than 155 regions worldwide (Gurses et al., 2020). Hospitals and healthcare professionals got challenged, as health systems were not well prepared for such an erratic and intricate event (Murthy, Gomersall, & Fowler, 2020). Suddenly, adjustments to healthcare practices were needed to ensure the safety of all patients, whether they were infected by the coronavirus or had other reasons to seek medical care. The coronavirus case sketches a typical situation in daily healthcare, as it was completely unpredictable and complex (Staender, 2015). The government and healthcare professionals were more than ever focusing on patient safety, by looking for the right preparatory actions and measures to manage this pandemic (Wilder-Smith & Freedman, 2020). The arising question is how caregivers are going to look back on this situation when most of the virus has been eradicated. Will they learn from the mistakes or will they learn from what went well?

For years, many sectors have looked at safety as the absence of adverse outcomes (Hollnagel, Wears, & Braithwaite, 2015). This view was developed between the 60’s and 80’s in the industrial sector and was later adopted by other sectors such as aviation and healthcare (Gill, 2004; Hollnagel et al., 2015). Safety was described as the occurrence of incidents, the identification of the causes and the presumed elimination of these causes to restore safety (Holbrook, Prinzell III, Stewart, Smith, & Matthews, 2019). In other words, unintended events can be found and fixed (Braithwaite, Wears, & Hollnagel, 2015). This approach was labelled as ‘Safety I’ by Hollnagel in 2014. At that time, he also constructed another view that not merely focuses on the absence of failures but looks at the ability to make things go right (Braithwaite et al., 2015). This approach is now known as ‘Safety II’. Unlike Safety I, Safety II focuses on maximizing the number of successful events (McNab, Bowie, Morrison, & Ross, 2016). In this manner, the definition of safety is not only ‘to avoid things going wrong’, but also ‘to ensure things going right’.

5

2014). To cope with cost increases, standardization of care and detailed prescriptions are being used (Bal, Stoopendaal, & Van de Bovenkamp, 2015).

In sum, a new, more positive line of thought on patient safety is on the way, but its implementation is still mainly theoretical due to the lack of acknowledgement of practical implementation. Healthcare professionals should refine their negative mindset on safety by taking Safety II into account in addition to Safety I and consider the possible ways of implementing it. If the Safety II approach is taught and implemented well, it may lead to the change of increasing the patient safety of healthcare systems through a better reflection by healthcare professionals on how day-to-day accomplishments are achieved in the complex healthcare sector (McNab et al., 2016).

By performing a systematic literature review, this study aims to contribute an overview of the existing literature on the implementation of Safety II. This overview provides results of significant studies that show how Safety II has been adopted and can therefore help healthcare professionals and organizations to further explore the topic of resilience, Safety I and Safety II. They benefit from this study by gaining knowledge about the implementation of Safety II in order to explore the possibilities of transitioning to resilient healthcare. Based on the proposal in the introduction, the following research question is formulated:

Which insights can be found in existing healthcare literature to stimulate healthcare professionals to implement resilience in patient safety?

This study is of great social significance, as it indirectly applies to every individual seeking medical care, and the caregivers providing this medical care. It is clear that all patients call for a high level of patient safety. This study is also important for researchers and policy makers, as they finally embrace the patient safety agenda due to the fact that research and policy in terms of patient safety have recently been influenced by the medical profession (Mannion & Braithwaite, 2017). The goal of this study is to show the value of Safety II and the possibilities for healthcare professionals and organizations to start thinking about patient safety in a positive sense. The findings may serve as a guidance for organizations and healthcare providers to support the development of resilience skills. Furthermore, this review can contribute as a preliminary study for future research that explore the field of resilience and Safety II.

6

Background

Safety II is specifically written with Roman numerals, because by definition it includes Safety I (ZonMw, 2019). This comprehensive theory originates from the concept of ‘Resilience Engineering’ (RE), which requires a different view of safety than what it has always been. This also implies another way of applying existing methods (Hollnagel, 2014). Safety II - and thus RE - assumes that people can adjust their performance to their environment and its conditions, since humans are the recourse for resilience and flexibility. It is believed that this is the reason why systems work (Hollnagel, 2014). RE recognizes the complexity of systems and understands that the linear cause-effect chains described by Safety I are limiting the performance of the healthcare sector (Braithwaite et al., 2015; Hollnagel et al., 2015). Although traditional safety strategies should still be used to manage the results of predictable situations, organizations must also acknowledge resilience in order to restore unpredictable events by quickly recovering and adapting (Bastan et al., 2018). Only when an organization can anticipate risk signals, and deal with, recover, and learn from incidents, can it successfully cope with high-risk processes (Nemeth, Wears, Woods, Hollnagel, & Cook, 2008).

According to Hollnagel (2017), resilience requires three kinds of ability: ‘‘An ability to adjust performance to the conditions, an ability to respond to changes, disturbances and opportunities, and an ability to do so in a flexible and timely manner’’ (p.26). Resilience also implies that these abilities emerge before, during and after a certain event happens (Hollnagel, 2017). From this it can be concluded that the practices of Safety II can take various forms. A well-known example is the FRAM method that was developed, also by Hollnagel (2012), to focus on everyday activities instead of the potential errors. The FRAM provides an insight into how it is possible to perform safety analyses without splitting systems into different parts and being independent on the concept of causality (Hollnagel, 2012). Another, rather new, health-related method is the Resilience Engineering Tool to Improve Patient Safety (RETIPS), which aims at obtaining everyday cases that can underpin resilience. It emphasizes the success of healthcare professionals, in order to deliver high quality care and keep patients safe in spite of the unpredictable and complex healthcare context (Hedge et al., 2020). These methods are foreground examples, but there might be other approaches in which Safety II can be found. For instance, thoughts can be given to performance support before or a debriefing after a certain intervention (Dieckmann et al., 2017; Karsh, Holden, Alper, & Or, 2006).

Work-as-Done and Work-as-Imagined

7

this (Mannion & Braithwaite, 2017). However, frontline workers such as physicians and nurses truly know how the work is done on a daily basis in the complex and varied healthcare setting (Braithwaite et al., 2015; Mannion & Braithwaite, 2017). Hence, work-as-imagined and work-as-done always differ from each other, and this gap gets wider as soon as people stand further away from the actual actions on the frontline (Braithwaite et al., 2015). Safety II enables work-as-done and work-as-imagined to be reconciled, and that the focus lies on what needs to be done to deliver safe care. Nevertheless, there are counterarguments as to whether or not Safety II is actually some sort of work-as-imagined. Smith and Valenta (2018) argue that the agreed core competencies of the workforce are more important than the approach (i.e. Safety II).

Implementation

The view on errors - Safety I - has contributed to the systematic approach to safety. However, as with the concerns about Safety I, there are concerns with this standardization. This is because it uses a lot of capacity, as a lot of standards are being developed by organizations and the government. Also, these standards do not take local variability into account, which impedes the desired transparency (Bal et al., 2015). Another approach that is more related to Safety II is to look at the cost-effectiveness of patient safety interventions, for example simulation-based training. Simulated training allows the healthcare professional to learn in a recreated, yet realistic environment where he can train to be resilient by learning from both his mistakes and successes. As a result, the patient is removed from the learning curve, which consequently improves the patient safety (Aggarwal et al., 2010). Further, Pederson (2016) argues that the approach to safely deliver care is by focusing on both the current safety system and the caregiver’s safety habits because the healthcare’s components, stability and change, are inseparable and connected.

To put the concept of Safety II into practice, several Dutch federations in the healthcare sector have come together in 2018 to develop a new strategy for patient safety (Leistikow & Bal, 2020). One of the three pillars of this national strategy focuses on Safety II, which aims to set up a learning program that will initiate a movement that will further embed the safeguarding of the approach in practice. The research program stimulates research into effective elements of the Safety II approach in the practice of Dutch hospitals. By the time of 2023, the organizations hope that it has given a positive direction to thoughts about healthcare (ZonMw, 2019).

Method

8

question. Due to the research protocol, the readers can readily assess its completeness and repeatability (Righi, Saurin, & Wachs, 2015).

This section first discusses the keywords and databases that have been chosen to cover a broad, yet specific research area to make sure that all significant aspects are included. Second, the search strategies are elaborated. Third, the selection criteria are identified and followed by an explanation of the data analysis.

Keywords

For a systematic literature review, it is important to explore all terms encompassing the research question in order to cover as large a research area as possible, in order for it to be a comprehensive study. This means that synonyms and related words must also be searched for and used as keywords, which applies to terms as resilience and healthcare. For the terms patient safety and implementation there is not an often used synonym, so there is no need to search for it. However, conjugations can be used in the existing literature, so it is wise to search for these as well, using ‘implementing’ as an example. This can be done by placing an asterisk (*) behind the term, so that the search engine also searches for the conjugations. In addition, placing a word between quotation marks (‘‘ ’’) ensures that this entire expression is searched for in the search engine.

The aforementioned instructions lead to the following terms of healthcare: healthcare, ‘‘health care’’, hospital*, and clinic*. For resilience, the following terms are used: resilien*, ‘‘Safety II’’, ‘‘work-as-imaged’’ and ‘‘work-as-done’’. The latter two are regarded as a combination, because they are also used inseparably in literature. As mentioned before, no synonyms for patient safety have been identified, so this term is only used in this manner. This also applies to the terms performance support and debriefing, which are meant to find more practical examples. Furthermore, to avoid an overload of keywords, only the search terms implement* and debrief* are used.

Databases

Three search engines were chosen based on their relevance to the subject of this study. Availability of the search engines for the researcher is also taken into consideration. The following search engines have been selected: ‘EBSCO’, ‘PubMed’, and ‘Web of Sciences’. These search engines use various databases, all of which were included in the research.

Search Strategies

There is a total of twenty search strategies to ensure that no articles are omitted in the search. The objective is to find articles on the implementation of Safety II on patient safety in healthcare, so these search terms are used as follows:

1. ‘‘patient safety’’ + implement* + healthcare + resilien* 2. ‘‘patient safety’’ + implement* + healthcare + ‘‘safety II’’

9

4. ‘‘patient safety’’ + implement* + healthcare + ‘‘performance support’’ 5. ‘‘patient safety’’ + implement* + healthcare + debrief*

6. ‘‘patient safety’’ + implement* + ‘‘health care’’ + resilien* 7. ‘‘patient safety’’ + implement* + ‘‘health care’’ + ‘‘safety II’’

8. ‘‘patient safety’’ + implement* + ‘‘health care’’ + ‘‘work-as-imagined’’ + ‘‘work-as-done’’ 9. ‘‘patient safety’’ + implement* + ‘‘health care’’ + ‘‘performance support’’

10. ‘‘patient safety’’ + implement* + ‘‘health care’’ + debrief* 11. ‘‘patient safety’’ + implement* + hospital* + resilien* 12. ‘‘patient safety’’ + implement* + hospital* + ‘‘safety II’’

13. ‘‘patient safety’’ + implement* + hospital* + ‘‘work-as-imagined’’ + ‘‘work-as-done’’ 14. ‘‘patient safety’’ + implement* + hospital* + ‘‘performance support’’

15. ‘‘patient safety’’ + implement* + hospital* + debrief* 16. ‘‘patient safety’’ + implement* + clinic* + resilien* 17. ‘‘patient safety’’ + implement* + clinic* + ‘‘safety II’’

18. ‘‘patient safety’’ + implement* + clinic* + ‘‘work-as-imagined’’ + ‘‘work-as-done’’ 19. ‘‘patient safety’’ + implement* + clinic* + ‘‘performance support’’

20. ‘‘patient safety’’ + implement* + clinic* + debrief*

These search strategies can be used in the same way in the three search engines. However, there are a few differences in search options. First, in EBSCO the first three search terms are searched for in the search field Abstract. The search field Title has been used for the fourth and when possible for the fifth term. Second, the search field Title/Abstract in PubMed was used for the first three terms and Title was used for the fourth and possibly fifth term. Last, in the Web of Sciences the search field Topic was used for the first three search terms, which includes the search in the titles, abstracts, authors keywords and Keywords Plus. Again, the fourth and possibly fifth search terms are only searched for in titles.

For all searches in the databases an AND operator was applied between the search concepts, except between the terms work-as-done and work-as-imagined. An OR operator was applied between these two terms, so it retrieved results containing one of these two search terms. No other limitations have been applied in the searches.

Selection Criteria

10 Data Analysis

After the selection procedure, the remaining articles have been assessed on their quality. Since both primary and secondary data was used, two different methods were used for this assessment. Primary data was assessed using the Standard Quality Assessment Criteria for Evaluating Primary Research Papers (Kmet, Lee, & Cook, 2004). This is done by means of scoring systems, in which multiple items that meet the specific criteria are scored at 2, 1 or 0, which implies that the criteria are fully, partly, or not addressed, respectively. As indicated by the method, different questionnaires are used for quantitative and qualitative research, and in mixed methods research both questionnaires were used. For some questions it was possible that they were not applicable to the study, so these were answered as ‘N/A’ and were not included in the total score. After going through this process, the authors recommend a relatively liberal cut-point of 50% of the total score (Kmet et al., 2004). This remained articles that meet the quality requirements.

No comprehensive criteria could be found for assessing articles based on secondary data, so this had to be devised. This was done by combining the checklist for evaluating secondary data of Stewart and Kamins (1993) and checklists found on university websites intended for their students to evaluate secondary data. This resulted in a total of seven questions with which the articles based on secondary data were assessed. These questions can be found in Table 5 of Appendix A. The remaining articles after the quality assessment could proceed to the following analysis.

The next stage of the research consisted of an in-depth analysis of the data of the articles that survived the selection process. To analyze the data, first the data was collected concerning the year of publication, country/region of data collection, main objective of the study, type of research used (i.e. qualitative, quantitative, mixed methods), and the manner of data collection. An overview of this can be found in Table 1.

Subsequently, content analysis was conducted to classify the data from the relevant articles into different categories. Content analysis is a technique to restrict large bodies of text into content categories (Stemler, 2000), which can be linked to each other to identify dominant findings (Dada, 2018). It also makes it possible to synthesize previous research in a transparently and scientifically based manner (Tranfield, Denyer, & Smart, 2003). Therefore, this data analysis is appropriate as a method for this study. The content analysis in this study is based on different manners of implementing resilience to benefit patient safety, which are classified into different categories. The result of the abovementioned data analysis is an overview of various types of implementation of Safety II in the health sector, which is presented in the next section.

Results

11 Search Results

Applying the twenty search strategies in the three search engines resulted in a total of 202 articles. The searches were carried out on 5 June 2020. In the resulting articles, duplicates were found from both the same or the other search engine. After these duplicates were manually removed, the number of articles was reduced to 55. Next, the remaining articles were screened on the predefined selection criteria, which excluded twelve articles. This assessment left 43 articles meeting the selection criteria. The next step was to assess the articles on their title and abstract. In case the relevance of the article for this analysis could not be clearly determined from the title or abstract, the article content was screened. This led to a further reduction of 21 articles to 22 remaining articles. By reviewing the references of these articles, ten relevant articles which brought in new insights have been added to the total number of articles, resulting in 32 articles. The next step was the quality assessment, the results of which can be found in Appendix A. This assessment resulted in ten articles that did not meet the minimum quality standards. Therefore, in the end, a total of 22 number of articles remained for further in-depth analysis. A graphical representation of the article selection process is presented in a flow chart (Figure 1).

Included Studies

12 EBSCO: 82 potentially relevant articles PubMed: 35 potentially relevant articles Web of Science: 85

potentially relevant articles

Total of 202 potentially

relevant articles

Total of 55 articles for

screening

Total of 43 for further

assessment Total of 22 studies 147 duplicates excluded 12 excluded based on selection criteria 21 excluded based on

title and abstract

Total of 32 studies before

quality check

10 included based on

search of reference lists

Total of 22 included

studies

10 excluded based on

quality assessment

Table 1. Included studies # Author &

publication year

Country/region Main objective of study Type of research

Data collection

2 Back et al., 2017 United Kingdom To examine escalation policy in theory and practice, using resilient

health care principles to identify opportunities for improving the way escalation is planned and managed.

Qualitative Conceptual analysis,

ethnography (observations, interviews), CARE

4 Blandford, Furniss, & Vincent, 2014

United Kingdom To raise awareness of the design and use of interactive medical

devices.

Literature review -

7 Catchpole & Alfred, 2018

USA To raise awareness of the application of the naturalistic

decision-making (NDM) paradigm that can help reveal clinical reasoning ‘as done.’ Literature review - 9 Clay-Williams, Hounsgaard, & Hollnagel, 2015 Australia, Denmark

To investigate whether FRAM can be used to identify process elements in a draft guideline that are likely to impede

implementation by conflicting with current ways of working.

Qualitative Case study,

interviews, FRAM

11 Dubois & Zedreck Gonzalez, 2018

USA To describe a practice change initiative implemented to enhance

resiliency in new nurses enrolled in a 1-year nurse residency program.

Quantitative Case study, survey

17 Hedge et al., 2020a USA To revise and validate the design of RETIPS. Qualitative Semi-structured

interviews, RETIPS

19 Kellogg & Fairbanks, 2018

USA To create awareness about the effect of fatigue and shift duration on

cognitive performance.

Literature review -

23 Patterson et al., 2016 USA To raise awareness of a loss of system resilience, due to increased

difficulty in performing macrocognition functions, associated with the implementation of new information technology.

Literature review -

36 Anderson et al., 2016 United Kingdom To test the feasibility of translating resilience engineering (RE)

concepts into practical methods to improve quality by designing, implementing and evaluating interventions based on RE theory.

Qualitative Ethnography

(observations, semi-structured

interviews), workshops, CARE

38 Damen et al., 2018 Australia,

Netherlands

To assess preoperative anticoagulation management in everyday practice and explore the usability and utility of FRAM.

Qualitative Semi-structured

14

40 Hegde et al., 2020b USA To elicit detailed information from frontline health care workers

regarding real-life examples of resilience.

Qualitative Semi-structured

interviews, RMF

41 Iflaifel et al., 2020 United Kingdom To identify how Resilient Health Care (RHC) is conceptualized,

described and interpreted in the published literature, to describe the methods used to study RHC, and to identify factors that develop RHC.

Literature review -

46 Magill et al., 2017 USA To routinely perform postoperative debriefs and evaluated the

impact of debriefing on operating room safety culture.

Quantitative Survey

49 Brindle et al., 2018 USA To identify and raise awareness of surgical debriefing programs to

improve system performance.

Qualitative Semi-structured

interviews

53 Mullan et al., 2013 USA To determine if debriefing of resuscitations to improve clinical

performance.

Quantitative Observations,

DISCERN

57 Hart, Brannan, & De Chesnay, 2014

USA To describe nursing research that has been conducted to understand

the phenomenon of resilience in nurses.

Literature review -

58 McNab et al., 2016 United Kingdom To create understanding of patient safety performance and

educational needs using the Safety-II approach for complex systems

Literature review -

59 Hollnagel, 2013 United Kingdom To explain the Resilience Analysis Grid (RAG) that can determine

how well a system does on the four basic abilities that are necessary for resilient performance.

Literature review -

60 Raben et al., 2018 Denmark To develop a generic method to identify leading indicators (MILI)

in healthcare.

Qualitative Case study, FRAM,

interviews

61 Hegde et al., 2014 USA To develop a Resilience Engineering Tool to Improve Patient

Safety (RETIPS) that can be implemented organization-wide for reporting and analysis of resilience-based cases

Qualitative Interviews

63 Wachs et al., 2016 Brazil, USA To discover where resilience skills (RSs) come from. Qualitative Case study,

interviews, survey, observations

64 Merandi et al., 2018 USA To identify how Safety II contributes in the high-complexity and

high-risk PICU to its remarkably low adverse drug event rate.

Findings

Content analysis was used to extract the data from the filtered articles. This way of analysis makes it possible to collect the data from the text and to categorize it. To analyze the findings, three categories have been developed that together contain the information required to draw conclusions. The categories are the A) resilience-oriented models, B) reason for implementation, and C) facilitators and barriers to implementation. All articles have been analyzed with these three categories in mind, so each finding could be classified in a specific category. The findings have been placed in one of the three categories, after which subcategories have been developed to structure these findings more clearly. Data outside the scope of this study was not included in the analysis or categories. The following paragraphs describe the categories and their subcategories separately.

Category A – Resilience-oriented models. In several articles, different methods or models were described or used which are based on resilience. Nine of these have been identified in eighteen articles, which can be found in Table 2. Next, the subcategories are discussed separately.

Table 2. Findings in Category A

A1: Resilience Engineering. The concept of Resilience Engineering (RE) allows us to understand that healthcare systems are complex and adaptive (Anderson et al., 2016; Back et al., 2017; Hollnagel, 2011), and are characterized by variability and unpredictability (Back et al., 2017; Iflaifel, Lim, Ryan, & Crowley, 2020). This is also the reason that predefined protocols do not always fit the everyday clinical work (Back et al., 2017; McNab et al., 2016), because ‘performance varies as situations vary’ (Hedge et al., 2014, p.803). RE tries to close the gap between work-as-done and work-as-imagined, especially between healthcare staff on the frontline and other professions such as analysts and policymakers (Hegde et al., 2020b; Kellogg & Fairbanks, 2018). It was thought that the traditional approaches in the safety improvement were limiting its success, but RE had a major influence on this view (Anderson et al., 2016). RE is central to the Safety II paradigm, in which adaptation to the variability is an essential feature in everyday clinical work (Hegde et al., 2020a; McNab et al., 2016).

General findings Article numbers

A1 Resilience Engineering (RE) 2/17/19/36/40/41/58/59/61/63

A2 Concepts for Applying Resilience Engineering (CARE) 2/36/41

A3 Functional Resonance Analysis Method (FRAM) 9/38/41/58/59/60

A4 Crew Resource Management (CRM) 4/7/49

A5 Naturalistic Decision Making (NDM) 7/23

A6 Resilience Engineering Tool to Improve Patient Safety (RETIPS) 7/41/61

A7 Resilience Analysis Grid (RAG) 41/59/61

A8 Resilience Mapping Framework (RMF) 40/41

A9 Debriefing In Situ Conversation after Emergent Resuscitation Now

(DISCERN)

16

This gives a comprehensive view in which learning from successful events (i.e. Safety II) complements learning from incidents (i.e. Safety I; Hedge et al., 2014; Iflaifel et al., 2020; McNab et al., 2016). RE enhances the system’s resilience in both the expected and the unexpected, adverse situations (Hegde et al., 2020b; Wachs, Saurin, Righi, & Wears, 2016).

A2: Concepts for Applying Resilience Engineering. In 2016, the CARE was developed by Anderson and colleagues (Iflaifel et al., 2020). This model provides a framework to examine how the resilience of healthcare organizations manifests itself, how it contributes to events and how it can be strengthened. The model does not focus on the full complexity of healthcare systems but looks at theoretical aspects that can be studied (Anderson et al., 2016). Back et al. (2017) used the CARE model to explore that healthcare staff must adapt their activities due the misalignments between demand and capacity, so that they can still achieve the desired results through their performance.

A3: Functional Resonance Analysis Method. FRAM is a method in which models can be developed that clearly shows the differences between work-as-done and work-as-imagined in specific contexts, so the performance and variability on everyday clinical work can be understood (Clay-Williams, Hounsgaard, & Hollnagel, 2015; Damen et al., 2018; Iflaifel et al., 2020; McNab et al., 2016; Raben et al., 2018). A FRAM model consists of the performance’s activities called ‘functions’ and the dependencies among each other (Hollnagel, 2011; Raben et al., 2018), which are placed in one of six aspects: input, output, precondition, resources, control, and time (Clay-Williams et al., 2015). Such a performance can be visualized as a hexagon with six different labels (Damen et al., 2018). Due to the visual representation that can be built in the software tool called FRAM Model Visualizer, the FRAM is easily understandable (Clay-Williams et al., 2015). Methods such as ethnography and interviews can be used to generate data for producing a FRAM model (Clay-Williams et al., 2015). However, Iflaifel et al. (2020) found that it is necessary to use quantitative data to be able to assess the distribution of variability. The resulting, context-specific model differentiates work-as-done and work-as-imagined, allowing it to provide practical information on how a new procedure can work in the real setting where work is done as it is done (Clay-Williams et al., 2015). A special component of a FRAM model is that it can be used to learn more about how the work is done even before an incident occurs (Damen et al., 2018).

17

A5: Naturalistic Decision Making. Clinical decision making (CDM) corresponds with linear, process-oriented tasks with clear goals, but which does not fit with complex adaptive systems (Catchpole & Alfred, 2018). In contrast, naturalistic decision making (NDM) points out the human role, as it is used by healthcare professionals in everyday, unpredictable situations by performing cognitively complex tasks (Patterson et al., 2016). Catchpole and Alfred (2016) claim that ‘‘if CDM is clinical reasoning ‘as imagined’, then NDM is clinical reasoning ‘as done’ ’’ (p.224). Resilience is embedded in NDM, as it enables the healthcare professionals to work on their own instead of thoughtlessly following standard working procedures (Patterson et al., 2016). However, NDM and CDM are not mutually exclusive, since CRM establishes standard criteria while NDM emphasizes the flexibility that deviates from that standard (Catchpole & Alfred, 2018).

A6: Resilience Engineering Tool to Improve Patient Safety. This approach was developed with the idea that healthcare professionals keep patients safe more often than not in the complex and variable healthcare systems. The aim of RETIPS is to identify examples in daily care that can be considered as supporting resilience (Hedge et al., 2020a), and to create a comprehensive picture of how the work is actually done (Hedge et al., 2014). It therefore suits Safety II, as it makes learning from successes more practical (Iflaifel et al., 2020). The tool consists of questions that bring out examples of specific challenging situations to situations that are so everyday that they can hardly be called a specific situation – which may even be more important (Hedge et al., 2020a). A pilot of an anesthesia-resident version of RETIPS was implemented, showing that the tool has good potential to learning in a proactive and appreciative manner, expanding the view of learning from only unwanted outcomes (Hedge et al., 2020a).

A7: Resilience Analysis Grid. Hollnagel (2011) discussed that resilience is not something that a system has or is, but that it should be considered what enables resilient performance. He states that four abilities are the basis for this resilient performance: the ability to respond, to monitor, to learn and to anticipate. These four abilities can be assessed in order to determine whether a system is working properly by using precise questions, which together constitute the RAG. Therefore, RAG aims to identify the characteristic of a system in order to manage this system and especially ‘‘to develop its potential for resilient performance’’ (Hollnagel, 2011, p.11). Iflaifel et al. (2020) found that four studies used RAG as the basis for their studies. Additionally, the beforementioned RETIPS used RAG as a guidance tool for the development of a resilience profile of a system (Hedge et al., 2014).

18

makes it possible to narrow the gap between WAI and WAD (Hegde et al., 2020b). A practical example of the usage of the RMF is in the reflection of the work demands of anesthetists, and the identification of contrast between routine and non-routine aspects (Iflaifel et al., 2020).

A9: Debriefing In Situ Conversation after Emergent Resuscitation Now. As much as standardized protocols are avoided in RE, there is a need for a predefined protocol for debriefing. DISCERN is a standardized and structured checklist form to facilitate debriefing, which strives to be readily accessible and to be used quickly and easily. The tool asks, among other things, what went well and what could have gone better. This creates a positive direction for learning from both mistakes and successes (Mullan et al., 2013).

Category B – Reason of implementation. Organizations, healthcare professionals and researchers all have different reasons to implement resilience in their work. Four reasons have been identified as the main objectives to implement resilience, which were which are classified into four subcategories. The data was found in all 22 articles. The subcategories can be found in Table 3, which also contains the article numbers in which the data was found. The findings are discussed separately in the next part of this chapter.

Table 3. Findings in Category B

General findings Article numbers

B1 To bridge the gap between WAD and WAI 2/4/7/19/36/38/58

B2 To gain insight in variability and performance adaptation 7/40/58/60/61/64

B3 To develop a new view on safety 7/17/23/41/59/60

B4 To learn lessons for future implementation or studies 2/9/11/17/36/38/40/46/49/53/57/

60/63/64

B1: To bridge the gap between WAD and WAI. Better alignment between practice and policy can be reached trough understanding the gap between WAD and WAI (Back et al., 2017). The benefit from understanding the gap between these views is that the clinical processes are better understood (Catchpole & Alfred, 2018). Anderson et al. (2018) developed the CARE model to differentiate WAD and WAI in order for better understanding the gap. Kellogg and Fairbanks (2018) stated that leaders often do not take the circumstances, relationships, and resources of the people working on the frontline into account. Additionally, Damen et al. (2018) examined WAI and WAD to gain more profound insights in medication management and Blandford et al. (2014) explored the design and usage of medical devices to ensure that these are fit for purpose in practice. McNab et al. (2016) state that the next step is to bridge the gap between WAI and WAD by involving frontline workers into the analysis of WAD, in order to move towards Safety II.

19

sectors. Analyzing variability in healthcare can help to foresee adverse events, but it can also illustrate how expected events succeed despite this variability (Raben et al., 2018). Merandi et al. (2018) found that when reduction of unintended variation plays a major role in an organization, intended variation may be necessary and is therefore also a characteristic of Safety II. This is because of the paradox that variation can lead to both failure and successes (Hedge et al., 2014). Furthermore, possible variations of conditions can be handled by adaptation, which is needed to react to complexity (McNab et al., 2016). According to Hedge et al. (2020b), there are also 'underground adaptations' that 'quietly' contribute to the safety of care, but which can also increase system complexity. Understanding the complexity and adaptations is necessary for a successful approach to improve the quality of safety, which is in line with what Safety II suggests (McNab et al., 2016).

B3: To develop a new view on safety. Safety has traditionally been seen as reducing adverse events, which is now thought to be insufficient to increase resilience in existing systems (Hollnagel, 2011; Patterson et al., 2016; Raben et al., 2018). The traditional view of safety led to ‘command-and-control’ approaches, saw variability as something that could be managed, looked more at work-as-imagined instead of how the work was really done, and failed to see how people can cope with everyday complex clinical work (Catchpole & Alfred, 2018). However, healthcare professionals more often than not successfully keep patients safe (Hedge et al., 2020a). It is suggested that by focusing on these positive outcomes, new prospects of healthcare processes can be examined (Raben et al., 2018). It has been shown that health care staff who are exposed to new perspectives on the patient's condition understand the patient's situation better, without exacerbating the previous performance, but reducing the risk of cognitive bias (Iflaifel et al., 2020).

20

medication management research, and Clay-Williams et al. (2015) also used this methods to understand the feasibility of a new procedure.

Category C – Facilitators and barriers. This category presents the facilitators and barriers to implementation of resilience by healthcare professionals. Seven aspects have been identified that can work as facilitators, but, if not managed properly, can also be experienced as barriers. All results were found in all 22 articles. Table 4 lists the facilitators and barriers, and the article numbers in which the data was found. The subcategories are now addressed separately.

Table 4. Findings in Category C

General findings Article numbers

C1 Organizations and guidelines must help the healthcare professionals

to explore RE and Safety II

2/9/41/46/49/57/58/61/63

C2 Education and training help understanding the concept of resilience 2/11/40/41/46/49/57/58

C3 Healthcare technology and tools can be beneficial if it assists the work of healthcare professionals.

4/7/19/23/40/41/46/53

C4 Methods must be concise and easy to understand 38/53

C5 Healthcare workers must be in good wellbeing to be resilient 11/17/41/53/57/58/61/63/64

C6 There must be an understanding of the gap between

work-as-imagined and work-as-done

2/4/29/36/40

C7 Staff and systems need to shift away from the traditional approach to safety

4/17/19/58/59/61

21

al., 2016). In contrast, organizations can also learn about adaptability from their own healthcare staff (Back et al., 2017).

C2: Education and training help understanding the concept of resilience. Resilience can be taught through training and education in specific behaviors, thoughts, and actions (Dubois & Zedreck Gonzalez, 2018; Hart et al., 2014). Communication (Brindle et al., 2018; Dubois & Zedreck Gonzalez, 2018), adaptive capacity (Back et al., 2017), teamwork and leadership (Brindle et al., 2018; Hedge et al., 2020b; Iflaifel et al., 2020), and coping with stress (Dubois & Zedreck Gonzalez, 2018) can all be taught, so that it can have a positive impact on staff and patient satisfaction and safety in later times (Dubois & Zedreck Gonzalez, 2018: McNab et al., 2016). Resilience can also be developed through simulation training and debriefing tools, which teach both experienced and new healthcare professionals to deal with expected and unexpected situations in an adaptable way (Iflaifel et al., 2020). The debriefing might be an effective tool in education as it includes giving and receiving feedback (Magill et al., 2017). C3: Healthcare technology and tools can be beneficial if it assists the work of healthcare professionals. Hedge et al. (2020b) argue that accessibility to technology and equipment, among other things, is a success factor for resilience. Healthcare technology is being used more and more in providing care. This means that interactive medical devices should be designed in such a way that they help to reduce the risk of mistakes made my healthcare staff; design and use must be in line with everyday clinical practices (Blandford et al., 2014). Devices are sometimes even used for interventions when they are not made for that at all, which eventually do not benefit the care delivery (Blandford et al., 2014). Electronic health records have pop-up warnings to inform healthcare staff about patient details, but these are not always supportive in the real clinical world (Catchpole & Alfred, 2018). Kellogg and Fairbanks (2018) argue that there are not many tools yet designed to help with the mental task list of healthcare staff to support cognitive work. However, Patterson et al. (2016) state that the implementation of new health information technology can be the source of the loss of system resilience, as the implementation cause the workflow to change. This means that the system design should be supporting resilience performance (Iflaifel et al., 2020). Furthermore, the debriefing tool can improve attitudes toward patient safety and communication between clinicians but missing a standardized protocol can be a barrier to debriefing (Magill et al., 2017; Mullan et al., 2013).

C4: Methods must be concise and easy to understand. If a method or tool is implemented, it should be accessible and could be filled out quickly. The fact that the majority of debriefings took less than or equal to ten minutes makes the feasibility of implementing DISCERN high. (Mullan et al., 2013). In the study by Damen et al. (2018), the FRAM method seems to be a promising tool, partly because the healthcare professionals understand the relevance, background, and design of the model well.

22

balance, and preventing dissonance in the workplace (Hart et al., 2014). Additionally, it is found that individual characteristics such as self-efficacy, self-image, optimism, adaptability, empathy, and skill recognition are related to resilience (Dubois & Zedreck Gonzalez, 2018; Hart et al., 2014; Hedge et al., 2014). Furthermore, it is found that good relationships among colleagues is crucial as it reduces psychological barriers (McNab et al., 2016; Merandi et al., 2018). Communication and coordination in teams often are described as important factors underlying resilient care processes (Hedge et al., 2020a; Iflaifel et al., 2020; Mullan et al., 2013; Wachs et al., 2016). Nonetheless, creating a work-life balance is also crucial for nurses to develop and sustain a feeling of resilience (Hart et al., 2014).

C6: There must be an understanding of the gap between imagined and work-as-done. Knowledge is an important factor to ensure safety (Hedge et al., 2020b). However, there is still a difference between established protocols and targets and the actual work (i.e. between WAI and WAD; Anderson et al., 2016). The actual care delivery has not yet been fully described in policies (Back et al., 2017). Kellogg and Fairbanks (2018) describe it as if leaders are trying to implement appealing plans, but that a miscalculation of reality has been made, as a result of which the plans are not achievable in the real environment. From a different point of view, Blandford et al. (2014) argue that the design of medical devices can already be poor, because their use is imagined differently by the designers than how it actually works in practice.

C7: Staff and systems need to shift away from the traditional approach to safety. People and systems can still be rooted in the traditional approach to safety, so that they still see it as avoiding adverse events, which creates limitations and can be counterproductive (Hedge et al., 2014; McNab et al., 2016). For instance, reporting systems' efficiency is limited due to the fact that these approaches focus on the events that go wrong (Hedge et al., 2020a). However, it is the people who are flexible and adaptive who ensure that things often go well in complex systems such as healthcare (Kellogg & Fairbanks, 2018). Also, Hollnagel (2011) argued that a system can only have a resilience performance if it concerns the four abilities (i.e. to respond, monitor, learn and anticipate). Furthermore, decisions about the configuration of health technology and its usage protocols must be made and reviewed with the aim of increasing the quality of patient safety (Blandford et al., 2014).

Discussion and Conclusion

23

No articles from before 2013 have been used, which means that the data is still relevant to use. Additionally, the majority of the articles were based on studies that were carried out in western countries. This could imply that the western world is at the forefront of Safety II thinking. Because the concept of Resilience Engineering in healthcare is still fairly new, it is likely that it has not yet had time to move to other countries. Future research should show whether safety practices according to the Safety II approach can also be carried out in other countries.

Resilience Engineering and Safety II are interconnected and associated with each other. Therefore, it has been difficult to focus solely on the implementation of Safety II approaches in the literature. Although Safety II has been extensively described in the literature, research on practical elements is still lacking (ZonMw, 2019). In addition, resilience is defined differently in the literature. It is argued that it is the ability to recover from and overcome adverse situations or that it is the ability to adapt to these situations (McAllister & McKinnon, 2009; Newman, 2003). Hollnagel (2011) stated that a system cannot be resilient, but that resilience is a characteristic of its performance. Furthermore, Hedge et al. (2014) determined that resilience could be diversified into resilience that occurs at an event and ‘everyday resilience’, which actually continues throughout the day to prevent adverse events. This resembles the assumption that there are different perspectives on what resilience actually is. This can be clarified by the argumentation that resilience involves adapting to different situations, which also means that resilience can emerge in different ways in changing forms, which is also found in research (Southwick et al., 2014).

The assumption of different perspectives might stem from the assumption that resilience can be viewed from different angles. Jeffcot et al. (2009) discussed that there are three interrelated levels: the individual, micro-organizational and macro-organizational level, and Iflaifel et al. (2020) described it as micro, meso and macro levels. If we use the prior approach, the findings in Category C can be classified into these levels. People at the top of the organization need to let go of the traditional view of patient safety and understand that there might be a gap between work-as-done and work-as-imagined in order to give guidance to healthcare providers. Guidelines, protocols, and technology should contribute to the development of resilience skills, as should education and training. At the team level, communication and coordination have proven to be important to strengthen resilience, as has leadership. Good relationships between colleagues also help to maintain resilience. At the individual level, it has been shown that the wellbeing of healthcare professionals is related to resilience. Personal characteristics, stress levels and work-life balance play a role in the development and maintenance of resilience performance. As a consequence of these different views, research may not always present the work as it is done in reality, because at each level different findings may lead to different views of WAD and WAI, which is also supported by Iflaifel et al. (2020).

24

resilience and positive adaptability. In order to provide the best possible care, the focus has traditionally been on the triple aim, in which the patient experience, population health and costs are considered, but nowadays it has been extended to the quadruple aim, which also includes the work life and wellbeing of healthcare professionals (Dubois & Zedreck Gonzalez, 2018; West, 2016). Furthermore, education and training in developing resilience performance (C2) often appeared in the articles. In this way, novel and experienced clinicians can learn how they can cope with variation in every work and to deal with it in an adaptive manner. This study revealed that programs including effective communication and teambuilding appear to be highly valued when learning about resilience. Found in this study and in other literature is that mindfulness can help building resilience, as it positively affects the sense of wellbeing (Anderson et al., 2016; Rogers, 2016). Finally, many included studies in this review implement the concepts of resilience in an organization or department to find out what does and does not work (B4). Lessons can be learned by testing and discovering how resilience fits into specific contexts.

The models and methods in Category A provide an overview of the possibilities to create resilience performance. It is actually not possible to implement the principle of resilience as a stand-alone element in an organization or team, because it arises from the human factors. From this point of view, the methods can only help to understand where the gap between WAD and WAI exists. Consistent with the literature (Clay-Williams et al., 2015), this review found that the methods are particularly useful for identifying what is needed to achieve resilience in practice. Those models can be used as a tool to initiate a change in the behavior of healthcare professionals. In addition, the response to situations differs per context, and both frontline workers and managers have to decide for themselves whether they need to adapt to the context or whether it is possible to adhere to the standardized protocols to ensure patient safety, which also appears from the literature (Pasquini, Pozzi, Save, & Sujan, 2011; Provan, Woods, Dekker, & Rae, 2020).

At the beginning of writing this review, the coronavirus expanded as a pandemic around the world. At the time, the question was whether healthcare systems and professionals could withstand such a large and unpredictable situation. Over time, we gained more knowledge about resilience and concept of Safety II, and learned more about the coronavirus. It turns out that the healthcare sector does have the resilience qualities to cope with such a large and unexpected event (Legido-Quigley et al., 2020; Polizzi, Lynn, & Perry, 2020). People will look back on this situation to what they have learned about patient safety, adaptability, and resilience when a complex event reappears in the future, because that will happen.

Implications

25

empirical study that in turn can contribute to the implementation of Safety II. The analysis of this research can serve as a guideline in future literature and field research.

This study provides a structured overview of different methods and models to enhance the resilience performance of healthcare workers. The facilitators and barriers are also known in order to be able to transition to Safety II. In this way, this review gives direction to healthcare professionals so that they can easily explore resilience and apply it to their practices in order to ensure better patient safety. Also, organizations and managers can learn from this review by recognizing the need to implement resilience in their systems and by supporting their staff to develop skills to strengthen their resilience performance. In doing so, the organizations can align with their personnel and systems, so that the work as they imagine it is as close as possible to the work that is actually done.

Research Limitations and Further Research

There are several limitations to this study. First, relevant data may have been missed. This may be because studies have not yet been published, because the search terms used were not comprehensive sufficiently, or because the search was limited to the English language. In order to counteract this limitation, research has been done in both primary and secondary data so that the search was carried out in a broad research field. However, other resilience-oriented methods and debriefing tools have been found, but they fell outside the range of articles in this systematic literature review. Second, for the articles based on primary data a standard checklist for quality assessment could be found, but for secondary data this was more difficult. Those that were found did not seem to be comprehensive. To counter this limitation, an attempt was made to put one together by combining two checklists of reliable sources. However, it is possible that other researchers may take different issues into account when evaluating secondary data papers. Third, most of the studies included in this review were carried out in the USA or United Kingdom, which means that more studies are needed to find out whether the results are applicable to other countries or regions. Fourth, the studies were conducted at specified professions (e.g. nurses) or healthcare departments (e.g. emergency departments). This means that the results of these studies may differ if they were conducted in different settings. Last, the results and discussion of this systematic literature review are based on the researcher's reasoning. It is possible that other researchers may have different interpretations and conclusions because they look at it from a different perspective.

26 References

Aggarwal, R., Mytton, O. T., Derbrew, M., Hananel, D., Heydenburg, M., Issenberg, B., ... & Ziv, A. (2010). Training and simulation for patient safety. BMJ Quality & Safety, 19(2), i34-i43.

Anderson, J. E., Ross, A. J., Back, J., Duncan, M., Snell, P., Walsh, K., & Jaye, P. (2016). Implementing resilience engineering for healthcare quality improvement using the CARE model: a feasibility study protocol. Pilot and feasibility studies, 2(61), 1-9

Back, J., Ross, A. J., Duncan, M. D., Jaye, P., Henderson, K., & Anderson, J. E. (2017). Emergency department escalation in theory and practice: a mixed-methods study using a model of organizational resilience. Annals of emergency medicine, 70(5), 659-671.

Bal, R., Stoopendaal, A. M., & van de Bovenkamp, H. (2015). Resilience and patient safety: how can health care regulations contribute?. Nederlands Tijdschrift voor Geneeskunde, 159, 1-4.

Bastan, O., Benesl, T., & Fiedler, P. (2018). Resiliency, the path to Safety II. IFAC-PapersOnLine,

51(6), 468-472.

Blandford, A., Furniss, D., & Vincent, C. (2014). Patient safety and interactive medical devices: realigning work as imagined and work as done. Clinical risk, 20(5), 107-110.

Braithwaite, J., Wears, R. L., & Hollnagel, E. (2015). Resilient health care: turning patient safety on its head. International Journal for Quality in Health Care, 27(5), 418-42.

Brindle, M. E., Henrich, N., Foster, A., Marks, S., Rose, M., Welsh, R., & Berry, W. (2018). Implementation of surgical debriefing programs in large health systems: an exploratory qualitative analysis. BMC health services research, 18(210), 1-14.

Catchpole, K., & Alfred, M. (2018). Industrial conceptualization of health care versus the naturalistic decision-making paradigm: Work as imagined versus work as done. Journal of Cognitive

Engineering and Decision Making, 12(3), 222-226.

Clay-Williams, R., Hounsgaard, J., & Hollnagel, E. (2015). Where the rubber meets the road: using FRAM to align work-as-imagined with work-as-done when implementing clinical guidelines. Implementation Science, 10(125), 1-8.

27

Dada, O. (2018). A model of entrepreneurial autonomy in franchised outlets: a systematic review of the empirical evidence. International Journal of Management Reviews, 20(2), 206-226.

Damen, N. L., Moesker, M. J., Braithwaite, J., Kaplan, J., Hamming, J. F., & Clay-Williams, R. (2018). Preoperative Anticoagulation Management in Everyday Clinical Practice: An International Comparative Analysis of Work-as-Done Using the Functional Resonance Analysis Method. Journal of Patient Safety, 1-9.

Dieckmann, P., Patterson, M., Lahlou, S., Mesman, J., Nyström, P., & Krage, R. (2017). Variation and adaptation: learning from success in patient safety-oriented simulation training. Advances in

Simulation, 2(1), 1-14.

DuBois, C. A., & Zedreck Gonzalez, J. F. (2018). Implementing a Resilience-Promoting Education Program for New Nursing Graduates. Journal for Nurses in Professional Development, 34(5), 263-269.

Eklöf, M., Törner, M., & Pousette, A. (2014). Organizational and social-psychological conditions in healthcare and their importance for patient and staff safety. A critical incident study among doctors and nurses. Safety Science, 70, 211-221.

Gill, G. K. (2004). Perception of safety, safety violation and improvement of safety in aviation: Findings of a pilot study. Journal of Air Transportation, 9(3), 43-55.

Gurses, A. P., Tschudy, M. M., McGrath-Morrow, S., Husain, A., Solomon, B. S., Gerohistodoulos, K. A., & Kim, J. M. (2020). Overcoming COVID-19: What can human factors and ergonomics offer?.

Journal of Patient Safety and Risk Management, (25)2, 49-54.

Hart, P. L., Brannan, J. D., & De Chesnay, M. (2014). Resilience in nurses: an integrative review. Journal of Nursing Management, 22(6), 720-734.

Hegde, S., Hettinger, A. Z., Fairbanks, R. J., Wreathall, J., Krevat, S. A., & Bisantz, A. M. (2020a). Knowledge Elicitation to Understand Resilience: A Method and Findings From a Health Care Case Study. Journal of Cognitive Engineering and Decision Making, 14(1), 75-95.

28

Hegde, S., Wreathall, J., Hettinger, A. Z., Fairbanks, R. J., Wears, R. L., & Bisantz, A. M. (2014). Towards the development of a resilience engineering tool to improve patient safety: the RETIPS approach. In Proceedings of the Human Factors and Ergonomics Society Annual Meeting,58(1), 803-807. Los Angeles, CA: SAGE Publications.

Holbrook, J., Prinzell III, L. J., Stewart, M. J., Smith, B. E., & Matthews, B. L. (2019). Resilience and safety for in-time monitoring, prediction, and mitigation of emergent risks in commercial aviation.

20th International Symposium on Aviation Psychology, 109-114.

Hollnagel, E. (2011). RAG - The resilience analysis grid. In Resilience engineering in practice. A

guidebook. Farnham, UK: Ashgate Publishing.

Hollnagel, E. (2012). FRAM, the Functional Resonance Analysis Method. Farnham, UK: Ashgate Publishing.

Hollnagel, E. (2014). Safety-I and Safety-II: The Past and Future of Safety Management. Farnham, UK: Ashgate Publishing.

Hollnagel, E. (2017). Safety-II in Practice. New York, NY: Routledge.

Hollnagel, E., Wears, R.L., & Braithwaite, J. (2015). From Safety I to Safety II: A white paper. The

Resilient Health Care Net: Published simultaneously by the University of Southern Denmark, University of Florida, USA, and Macquarie University, Australia.

Iflaifel, M., Lim, R. H., Ryan, K., & Crowley, C. (2020). Resilient Health Care: a systematic review of conceptualisations, study methods and factors that develop resilience. BMC health services

research. 20, 1-21.

Jeffcott, S. A., Ibrahim, J. E., & Cameron, P. A. (2009). Resilience in healthcare and clinical handover.

BMJ Quality & Safety, 18, 256-260.

Karsh, B. T., Holden, R. J., Alper, S. J., & Or, C. K. L. (2006). A human factors engineering paradigm for patient safety: Designing to support the performance of the healthcare professional. BMJ

Quality & Safety, 15(1), i59-i65.

Kellogg, K. M., & Fairbanks, R. J. (2018). Approaching Fatigue and Error in Emergency Medicine: Narrowing the Gap Between Work as Imagined and Work as Really Done. Annals of emergency

29

Kmet, L. M., Lee, R. C., & Cook, L. S. (2004). Standard quality assessment criteria for evaluating

primary research papers from a variety of fields. Alberta Heritage Foundation for Medical

Research.

Legido-Quigley, H., Asgari, N., Teo, Y. Y., Leung, G. M., Oshitani, H., Fukuda, K., ... & Heymann, D. (2020). Are high-performing health systems resilient against the COVID-19 epidemic?. The

Lancet, 395, 848-850.

Leistikow, I., & Bal, R. A. (2020). Resilience and regulation, an odd couple? Consequences of Safety-II on governmental regulation of healthcare quality. BMJ Quality & Safety, 1-4.

Magill, S. T., Wang, D. D., Rutledge, W. C., Lau, D., Berger, M. S., Sankaran, S., ... & Imershein, S. G. (2017). Changing operating room culture: implementation of a postoperative debrief and improved safety culture. World Neurosurgery, 107, 597-603.

Mannion, R., Braithwaite, J. (2017). False dawns and new horizons in patient safety research and practice. International Journal of Health Policy and Management, 6(12), 685–689.

McAllister, M., & McKinnon, J. (2009). The importance of teaching and learning resilience in the health disciplines: a critical review of the literature. Nurse Education Today, 29(4), 371-379.

McNab, D., Bowie, P., Morrison, J., & Ross, A. (2016). Understanding patient safety performance and educational needs using the ‘Safety-II’ approach for complex systems. Education for Primary Care,

27(6), 443-450.

Merandi, J., Vannatta, K., Davis, J. T., McClead, R. E., Brilli, R., & Bartman, T. (2018). Safety II behavior in a pediatric intensive care unit. Pediatrics, 141(6), 1-8.

Mullan, P. C., Wuestner, E., Kerr, T. D., Christopher, D. P., & Patel, B. (2013). Implementation of an in situ qualitative debriefing tool for resuscitations. Resuscitation, 84(7), 946-951.

Murthy, S., Gomersall, C. D., & Fowler, R. A. (2020). Care for critically ill patients with COVID-19.

Jama Insights, E1-E2.

Nemeth, C., Wears, R., Woods, D., Hollnagel, E., & Cook, R. (2008). Minding the gaps: creating resilience in health care. In Advances in Patient Safety: New Directions and Alternative

Approaches. Vol. 3: Performance and Tools. Rockville, MD: Agency for Healthcare Research and

Quality.

Newman, R. (2003). Providing direction on the road to resilience. Behavioral Health Management,

30

Pasquini, A., Pozzi, S., Save, L., & Sujan, M. A. (2011). Requisites for successful incident reporting in resilient organisations. In Resilience engineering in practice. A guidebook. Farnham, UK: Ashgate Publishing.

Patterson, E. S., Militello, L. G., Su, G., & Sarkar, U. (2016). Characterizing a naturalistic decision-making phenomenon: loss of system resilience associated with implementation of new technology. Journal of Cognitive Engineering and Decision Making, 10(3), 229-243.

Pedersen, K. Z. (2016). Standardisation or resilience? The paradox of stability and change in patient safety. Sociology of Health & Illness, 38(7), 1180-1193.

Polizzi, C., Lynn, S. J., & Perry, A. (2020). Stress and coping in the time of COVID-19: pathways to resilience and recovery. Clinical Neuropsychiatry, 17(2), 59-62.

Provan, D. J., Woods, D. D., Dekker, S. W., & Rae, A. J. (2020). Safety II professionals: how resilience engineering can transform safety practice. Reliability Engineering & System Safety, 195(106740), 1-14.

Raben, D. C., Bogh, S. B., Viskum, B., Mikkelsen, K. L., & Hollnagel, E. (2018). Learn from what goes right: a demonstration of a new systematic method for identification of leading indicators in healthcare. Reliability Engineering & System Safety, 169, 187-198.

Righi, A. W., Saurin, T. A., & Wachs, P. (2015). A systematic literature review of resilience engineering: Research areas and a research agenda proposal. Reliability Engineering & System

Safety, 141, 142-152.

Robertson, H. D., Elliott, A. M., Burton, C., Iversen, L., Murchie, P., Porteous, T., & Matheson, C. (2016). Resilience of primary healthcare professionals: a systematic review. British Journal of

General Practice, 66(647), e423-e433.

Rogers, D. (2016). Which educational interventions improve healthcare professionals’ resilience?. Medical Teacher, 38(12), 1236-1241.

Smith, K. M., & Valenta, A. L. (2018). Safety I to Safety II: A paradigm shift or more work as imagined?: Comment on "False dawns and new horizons in patient safety research and practice".

International Journal of Health Policy and Management, 7(7), 671-673.

Southwick, S. M., Bonanno, G. A., Masten, A. S., Panter-Brick, C., & Yehuda, R. (2014). Resilience definitions, theory, and challenges: interdisciplinary perspectives. European journal of

31

Staender, S. (2015). Safety-II and resilience: the way ahead in patient safety in anaesthesiology. Current

Opinion in Anesthesiology, 28(6), 735-739.

Stemler, S. (2000). An overview of content analysis. Practical Assessment, Research, and Evaluation,

7(1), 17.

Stewart, D. W., & Kamins, M. A. (1993). Applied Social Research Methods: Secondary research. Thousand Oaks, CA: SAGE Publications, Inc.

Tranfield, D., Denyer, D., & Smart, P. (2003). Towards a methodology for developing evidence‐ informed management knowledge by means of systematic review. British journal of management,

14(3), 207-222.

Wachs, P., Saurin, T. A., Righi, A. W., & Wears, R. L. (2016). Resilience skills as emergent phenomena: a study of emergency departments in Brazil and the United States. Applied Ergonomics, 56, 227-237.

West, C. P. (2016). Physician well-being: expanding the Triple Aim. Journal of general internal

medicine, 31(5), 458-459.