Learning from Excellence and Safety-II: reframing patient safety (Plunkett)

The prevailing approach to patient safety is to define safety as the absence of harm. In this paradigm, safety is considered to be a condition in which as few things as possible go wrong. Arguably, this definition is incomplete as it only considers safety from a deficit-based perspective.

Reframing is a cognitive exercise through which concepts are viewed from alternative perspectives. Reframing can be applied to patient safety, allowing safety to be considered from a strengths-based perspective: that is safety can be considered to be a condition whereas many things as possible go right. This is the basis of Safety-II and a number of other new approaches to safety, including Learning from Excellence (LfE). Viewing safety from a strengths-based perspective allows new insights to appear through examination of the conditions and characteristics of success, rather than those of failure. Strengths-based approaches are intended to be used as a complementary approach to the prevailing approach.

The aims of this chapter are to put forward a case for reframing safety; to review some of the limitations of the prevailing approach to safety; to introduce the concepts of Safety-II and LfE; to review relevant cognitive considerations related to safety; to provide examples of reframing; and to highlight other strengths-based approaches, including exnovation and positive deviance.

The case for reframing patient safety

The prevailing definition of safety is incomplete

The prevailing approach to improving patient safety is to identify and eliminate harm. This deficit-based approach is on the basis of a paradigm in which safety is defined as the absence of harm. Arguably, the definition is incomplete as it only defines the condition by what it isn’t, rather than by what it is. It reduces events in health care into two mutually exclusive states: safe and unsafe. In reality, events in health care occur across a wide spectrum from exceptionally poor to exceptionally good. The vast majority of events in this spectrum result in successful outcomes, yet the prevailing approach to safety compels all improvement efforts to be focused on the minority of events which lead to failure (i.e. harm or near misses).

Concentrating all efforts to improve safety on events leading to failure results in missed opportunities to learn from events leading to success. Inquiry into success can shed light on the positive aspects by which safety can be defined. Characteristics and properties of healthcare events and interactions which create safety come into view, and can be considered alongside those which lead to failure. Thus, strengths-based approaches are intended to be complementary to deficit-based approaches.

Limitations of the deficit-based approach to safety

Assumption of linearity

The deficit-based approach to safety adopts ‘find and fix’ methodologies, which aim to make systems safer by identifying harm, or potential harm, and then eliminating the causes. Flaws in a system are thus ‘patched’ or ‘repaired’. Methodologies to identify causation of harm in health care are typically based on approaches taken from other safety-critical industries. A common methodology is Root Cause Analysis (RCA), which aims to identify root causes and contributory factors to adverse incidents or episodes of harm. This method was not developed specifically for health care. Its application in complex systems such as health care may be flawed, in part due to its assumptions about linear causation of events.179,180 RCA requires complex systems to be decomposed to causal chains of events, aligned in a linear fashion, in order for investigators to identify why adverse events happened, and how they may be prevented in the future.

Application of this approach to health care may result in oversimplified representation of complex systems. Health care is a complex adaptive system (CAS), in which conditions and events rarely occur in a simple, predictable linear fashion.179 Indeed, many conditions and events in health care are emergent properties of the system, and thereby unpredictable.27

Work-as-imagined is not the same as work-as-done

The reality of work in a CAS is often significantly different from the protocols and guidelines on which work is designed. The latter are often developed away from the ‘sharp end’ of work, and don’t account for the highly variable conditions in which work is carried out. In the Safety-II literature, this protocolised and procedural depiction of work is known as work-as-imagined (WAI); and is contrasted with the real work carried out by front-line staff, known as work-as-done (WAD).181

The prevailing approach to investigating safety incidents typically seeks to identify deviations from protocol, by comparing WAD with WAI. Safety investigations often find that protocols and guidelines were not followed, and then assume that this deviation is the reason for the error or harm. In this paradigm, human error is frequently identified as a root cause or contributory factor of harm: human fallibility is considered a risk, and human performance variability is identified as something to be mitigated and reduced.182,183 It follows that recommendations from safety investigations are often designed to reduce the variability of human performance through the introduction of constraints and guardrails. If WAD is not properly understood, the addition of constraints and guardrails may paradoxically make work more difficult, and thus less safe.181

Analysis of successful work, on the other hand, reveals that WAD often deviates from WAI across the whole spectrum of work: that is in success as well as failure. Performance variability is often key to create safety, rather than creating harm.184 Therefore, WAD is often inherently different from WAI. Reconciling this difference is part of the aim of Safety-II and other strengths-based approaches to safety, such as LfE. To do this, proactive, prospective exploration of everyday work (i.e. WAD) is required, including examination of the approximate adjustments and performance variability required to create safety.

There are more opportunities to learn from success than failure

Learning from success is arguably easier than learning from failure in at least two ways: first, success is far more prevalent than failure, and therefore success presents more opportunities for learning than failure.181 Second, there are fewer paths to success than to failure: a ‘thing’ can go wrong in a greater number of ways than it can go right. Therefore, understanding the causes of rare failure may yield less actionable intelligence than understanding the causes of frequent success.

Cognitive considerations

Negativity bias

Learning from adverse events is of great importance, especially for the patients and families who have suffered consequential harm. It is arguably equally important to learn from positive events (e.g. unexpected positive outcome), yet the deficit-based approach to safety does not provide a means through which this can be done. This is in part due to an innate negativity bias,185 which is reflected in our (human) tendency to be drawn to the small part of the system in which adverse events occur. We are more sensitive to negative events than to positive ones of equivalent value: thus, errors and harms are simply more obvious to us than success at the other end of the spectrum. Our preoccupation with negativity in health care may be enhanced by the primary aim of health care itself – the intention of medicine is to diagnose and treat illness. This ‘diagnose and cure’ approach to medicine is analogous to the ‘find and fix’ approach used in the prevailing approach to safety.


An additional relevant feature of human cognition is habituation.186 Frequently occurring events become habituated in our cognition, and thus become less and less noticeable. The rare, serious adverse event is much more easily noticed against a background of everyday success. Like many features of innate cognition, this phenomenon is useful most of the time: there is simply too much data to notice everything! Through careful inquiry it is possible to temporarily overcome this phenomenon and understand what we take for granted in everyday, successful work. Thus, it is possible to shed light on insights into why things go right most of the time, and conversely, why things occasionally go wrong.


We have a tendency to find patterns and create oversimplified representation of complex systems (e.g. RCA).187 It is very hard to ‘see’ complexity in a CAS, and thus safety investigations run the risk of oversimplification of causal chains of events.180 Modelling systems are available to better illustrate the manifold interactions and interdependencies of multiple functions within a system – for example, the Functional Resonance Analysis Method (FRAM).188 While these models are still an oversimplification, they have the potential to create a significantly improved depiction of WAD.

These cognitive ‘programmes’ (e.g. negativity bias, habituation, oversimplification) can be temporarily suspended to recognise and learn from success in everyday work in health care; and thus, significantly open the aperture of the lens through which we can gain insights from the whole spectrum of work.

How to reframe

General approach

Reframing is an exercise in which concepts can be challenged and viewed from alternative perspectives. It creates an opportunity to scrutinise, and learn from, part of a system hitherto unstudied. Relevant examples of reframing from the patient safety domain are shown in Table 10. The table shows the contrasting, and complementary, lines of inquiry into the same issue, from different frames. This illustrates how a complementary approach can be taken to almost any issue, by revealing a different part of the system to study. In this way, the whole landscape of work, both successful and unsuccessful, can provide learning opportunities.

Table 10: Examples of reframing

Practical examples of reframing safety


In recognition of the limitations of the prevailing approach to safety, the concept of Safety-II was introduced by Hollnagel et al.181 They named the prevailing, deficit-based approach to safety, ‘Safety-I’. Safety-I was contrasted with the novel, complementary concept of ‘Safety-II’, in which safety is considered to be a condition in which as many things as possible go right.

The application of this theoretical concept to practical work is challenging, especially if viewed from within the conventional confines of the Safety-I mindset. Reconciling the difference between WAD and WAI is one of the main features of Safety-II. A different approach is required from the backward-looking, exception reporting approach typical of Safety-I. Safety-II methods, therefore, are typically prospective, front-line-based interventions, which focus on understanding how real work is done before recommending and implementing system adjustments.

Safety-II methods seek multiple perspectives on work to appreciate the degree of complexity, along with close participation of front-line workers to gain a good understanding of WAD. Examples include a project in which central venous catheter bundle adherence was improved by redesigning the bundle protocol with insights gained from an in-depth study of WAD.189 This is an example of combined Safety-I (identifying where bundle adherence was poor) and Safety-II (understanding how WAD actually happens, including variability and adjustments in real work) approaches, to achieve a goal of increased rate of success.

Other Safety-II methods include FRAM,188 which can be used to create a model of a complex system to share understanding of how the various steps (functions) in a process are related. Rather than creating a linear flow diagram, FRAM results in a map through which the interdependencies of multiple steps can be visualised. Vulnerable or impactful steps can then be identified and potentially modified. An example of the application of FRAM in healthcare safety is illustrated in a study of blood sampling by Pickup et al.,190 in which the authors created a FRAM model of blood sampling to illustrate how the process works successfully. Key steps were identified in which potential downstream consequences of variability were highlighted. Insights from this type of study can be used to add resilience to a complex system.

Learning from Excellence

LfE is a social movement in health care which started in 2014 in a single paediatric ICU.191 The initiative is based on a strengths-based philosophy with two main aims: first, to gain new insights about safety by identifying and studying excellence; and second, to provide formal positive feedback between staff following excellent practice.

LfE is a complementary approach to the prevailing approach to patient safety. It arose, in part, as a response to a perception of increasing negativity associated with the patient safety industry, for which there was growing concern that healthcare staff were suffering adverse psychological consequences (referred to as second victim). In the LfE model, ‘excellence’ is not defined a priori, since the initiative aims to capture excellence ‘in the wild’ as judged by front-line staff. Reports are filed by staff members using reporting systems that typically are juxtaposed to adverse incident reporting systems. In some centres, patients also have access to the system. Reports are forwarded privately, but not anonymously, to cited individuals or teams to provide positive feedback, to enhance learning and to improve morale and staff experience. Previous research has established a correlation between staff experience and performance of healthcare organisations, as measured by multiple performance indicators.192

Selected reports are investigated in more detail using appreciative inquiry (AI), a strengths-based inquiry method which originated in health care.193 Insights from these inquiries, which may include innovative practice, are shared with stakeholders and, if practicable, adjustments made to working conditions and systems. In addition to gaining useful insights into successful work, the LfE approach aims to reinforce positive interactions between colleagues by identifying and appreciating pro-social, positive behaviours in the workplace. This unique feature of LfE addresses an important aspect of the initiative, through which a wider organisational culture could be positively influenced, for example, by increasing psychological safety.

Practical example of Learning from Excellence approach

LfE can be easily introduced through a variety of open, voluntary reporting systems. A more focused application of the initiative can also be used to drive change in a quality improvement setting. In a recent proof of concept study, LfE was used to positively reinforce clinician behaviours related to antimicrobial stewardship in a paediatric ICU.194 Selected positive (i.e. successful) behaviours including prescribing practice and antimicrobial selection and administration were reinforced with LfE reporting and AI interviews over a period of 6 months. Rates of some positive behaviours improved throughout the study, and the primary aim of safe reduction in overall antimicrobial consumption was achieved.

Comparison of Safety-II and Learning from Excellence

A number of Safety-II methods have now been described. While LfE overlaps with Safety-II, the two philosophies also deviate in some respects. Table 11 illustrates some of the similarities and differences between the two philosophies.

Table 11: Safety-I, Safety-II and Learning from Excellence

Other strengths-based approaches

The current article is not as an exhaustive review of strengths-based approaches to patient safety. Other approaches include exnovation and positive deviance. Exnovation is a process through which ‘hidden competence’ can be unmasked.195 The usual method employed in exnovation is observation of practitioners, typically with video ethnography. The participants then review the observations and identify and share how safety is created in daily work. Thus, exnovation is a methodology for capturing successful WAD, and can therefore be employed as a Safety-II methodology. Positive deviance is an approach to identify solutions which already exist within a system or community, but have hitherto not been fully appreciated.196 Individual behaviours, practitioners or teams who are successful, but deviant in their practice (i.e. significantly different from their peers/community) are identified for the purpose of amplifying and spreading their successful practice more widely. This has been applied in several settings, including health care.


The impact of healthcare-associated harm on patients and their families is considerable, but progress on reducing rates of harm has been disappointing.197 The prevailing approach to patient safety may be inadequate, as it only considers safety from a deficit-based perspective. Reframing allows safety to be considered from a strengths-based perspective and opens the door to alternative methods and tools to improve safety. Strengths-based approaches, such as LfE, can be used to unmask the positive characteristics of safety, many of which are behavioural, cultural and relational. LfE provides a method to recognise, appreciate and reinforce these positive factors. Safety-II is a concept from which stems multiple novel methods to improve safety through reconciliation of WAD and WAI. The challenge for the future is to integrate both perspectives to provide a balanced, holistic approach to safety.

Please visit the following link to see this article in full on the EJA website