Common assumptions in RCA
The local reality
The investigation team displays a thorough understanding of the event through the rational presentation of information
The information gathered by the investigation team is partial and incomplete
The purpose of an investigation is to establish a reliable account of what happened and why it happened
There is no single authoritative account of an event as the analysis of what happened is influenced by the emerging mental models of the people involved and interpreted through the collective wisdom of the investigation team
The investigation team’s task is to demonstrate cause and effect relationships and develop corrective actions that address each root cause or contributing factor
The findings of an investigation team are tentative and recommendations need to be confirmed in the local setting because it is not possible to capture all possible consequences of an event or anticipate all future possible situations where a similar event may occur
The incident investigation system takes into consideration the concerns of frontline personnel and is a tool for learning through the dissemination of positive actions that reduce or eliminate vulnerabilities identified
The consequences of an event are related to subjective factors that operate deep within the workplace independent of rational statements in incident investigation reports. Therefore, all conclusions remain open to review and require ongoing dialogue in the workplace
Developing insight into the way complex human systems interact and making connections within perioperative environments requires a shift in mindset about the knowledge generated from incident investigations [19]. Techniques like root cause analysis originate from industrial settings where, typically, the contributory factors to a defect in a stable system can be attributed to a limited number of physical causes. The nature of clinical adverse events is such that it is near impossible or exceedingly rare to frame an investigation around a single procedure or device. Human error is even more problematic as it is hard to link individual actions to discrete properties of a broken system [6, 8]. Rather than argue for a best method of incident investigation, the chapter presents a number of related propositions that can be used to make decisions about the most appropriate combination of tools that will help make sense of an adverse event. In contrast to the assumptions of Newtonian rationality about the universal application of methods, the guiding premise is that an investigation team needs to understand an event within the context of the operating environment and adapt the selection of tools and techniques accordingly [22]. The approach represents a shift away from assuming that there are broken properties to fix, to shaping a perspective of the event that best fits the nature of the problematic situation, and directs sparse organizational attention and resources towards the methods of inquiry that will provide a useful explanation.
A Cautionary Word About Methods of Inquiry
Most of the frustration with adverse event data and the slow progress with making changes in response to incident investigations can be related to either relying too heavily on a particular investigation technique to draw conclusions, or to making incorrect assumptions about the purpose of an inquiry. The trajectory of a serious adverse event is unique and unlikely to occur again in exactly the same pattern. Meta-analyses of RCA report data make the assumption that common factors can be categorized and aggregated across multiple (often high risk) clinical adverse events without due regard for the contextual factors that were particular to individual cases on a given day in a specific perioperative setting with a particular surgical team [23]. Aggregated RCA data, consequently, has little predictive value for future adverse events. This is challenging for regulators, risk managers, and the public to appreciate. Separated from the original context of action, system level aggregations of event data become a “cumulative mess” through the multiplication of known causes and effects [18].
When external governing bodies make these assumptions they tend to work from a basic set of definitions for the purpose of systematically organizing the consequences of multiple adverse events into categories within a measurable body of knowledge (e.g., incident management and reporting systems). At a universal level, the questions posed by regulators relate to what can be known generally about adverse events and clinician’s performance.
In contrast, clinicians deal with everyday interactions in context and relate knowledge construction to the dynamic of particular situations [5]. In order to address what is known or unknown about the risks and vulnerabilities in the perioperative setting methods are needed that enable the discovery of previously unrecognized problems (e.g., failure modes effect analysis , fault tree analysis, and probabilistic risk assessment) [24]. Questions at a contextual level relate to gaining a better understanding of operational matters across a department. When making sense of an adverse event it is important to find out what was known about the particular problematic situation by the people involved. An incident investigation draws on the experience of people working on the frontline in the clinical setting in order to reconstruct the event.
In summary, the three different ways of knowing represent three basic approaches to constructing knowledge about adverse events:
- 1.
Knowledge as transferring data. Policy makers and regulators look for what is known generally, from aggregated reports,
- 2.
Knowledge as learning about systems. Perioperative suite quality and safety programs seek to discover what is unknown or better understand known risks, and
- 3.
Knowledge as an ongoing dynamic. Local incident investigation teams work with what is knowable about an event from the circulating information about everyday clinical actions.
The points of intersection in the diagram (Fig. 41.1) represent the current state of knowledge about actual or potential problems. In practice, knowledge varies from situation to situation, is highly context dependent, and mediated through a process of translation by multiple people at different levels across the organization (For a thorough analysis of knowledge transfer, see [25]. Cook and Woods [19], discuss the impact of resources and constraints on knowledge at the point of care delivery in complex systems ).
Fig. 41.1
Different inquiry methods produce different types of knowledge
Building a Body of Knowledge About Adverse Events
Although root cause analysis was introduced into health care for the investigation of serious clinical adverse events, it uses causal reasoning from stable categories to deduce what happened (i.e., root cause analysis works from categories of what is generally known to break an event into knowable parts). The view of the system is drawn from a process of deduction from known factors. Curiously, the root cause analysis checklist of questions used by the US Veteran’s Administration (and adopted by other countries such as Australia) is labeled as a set of human factors questions, but the method is unlike most human factors investigations due to a focus on identifying specific causation and applying fixed categories in relation to multiple aggregated events [23, 26]. Without knowing the consequences for the local clinical workplace, and how the people involved defined the situation, there can be no meaningful understanding of the event and the depth of analysis is limited [18, 19].
In terms of the way knowledge is produced, most adverse event investigations fit between two antithetical positions: either there is a specific “root” cause to find and sort into causal statements for corrective action according to a standardized hierarchy imported from other industries [9], or they opt for the alternative, that a clinical adverse event is the outcome of multiple contributory factors that are open to explanation from different perspectives particular to the complexity and context of the situation [15]. The separation of human factors in an event from system issues, under the label of “human error” is arbitrary, reflecting a misguided commitment to investigation methods adopted from engineering without regard for the interplay between expertise and situational constraints in complex clinical environments [4, 6, 9, 19]. The selection of method often says more about the purposes and philosophy of the investigation team and the sponsor of the investigation than the event itself [26]. The choice of response to an adverse event will to a large extent determine whether the investigation team seeks to replace broken components of a system, identify a barrier to prevent recurrence, consider the redesign of particular tasks, or to optimize workplace systems by developing a better understanding of what people do at the local level [27].
An Overreliance on Rational Analysis Paralyzes Local Knowledge
Top-down quality and safety processes have been implemented in all major health care systems for the management of adverse events. The situation in health care a decade ago was that decision makers needed to be mobilized to turn the idea of the patient safety movement into an organizational reality [28]. There is a growing body of literature that documents the implementation of the resultant processes such as root cause analysis (RCA) for the investigation of serious adverse events [5]. However, the assumptions about using retrospective approaches to locate patterns of error within health care systems need to be challenged due to an over reliance on rational analysis as a basis for understanding breakdowns in care delivery [29]. The initial implementation of safety improvement programs introduced structured processes for thinking about the causes and contributing factors to adverse events. As health departments and jurisdictions accumulated data about findings from RCAs assumptions were made about the transferability of what was known about past events from generalized data aggregated from multiple RCA reports [23]. Informal corridor conversations about care lack the apparent rigor of rational management sanctioned incident investigations. The inherent risk in the pursuit of more reliable adverse event data is the paralysis of knowledge transfer at the local level which is the most important level for developing an understanding of how people manage constraints and regain control of unexpected events [6, 30]. A philosophical commitment to the prevention of adverse events feeds into a belief that systems are generally consistent and reliable. The reality is rather different in complex clinical systems. Accepting that good people sometimes make poor judgments and decisions is more likely to lead to an understanding of the inconsistencies that are common in everyday human interaction with complex clinical and organizational systems [30].
In order to manage this dilemma it is necessary to consciously reflect on the models that the perioperative department selects to guide incident investigation [26]. If the department is primarily concerned with external reporting there is likely to be a focus on identifying the organizational factors related to adverse events. A limitation of this risk averse approach is that perioperative care directly depends on what humans do each day in the operating room environment where only the indirect impact of organizational decisions are seen. Health care is quite different to other industries and trying to identify a rational explanation for interaction in human systems can be problematic and over reaching. Clinical work involves a level of complexity not encountered in stable closed systems where incident investigation heuristics such as the “Swiss cheese model” originate [26, 31]. It is a constant challenge to resist the management imperative to produce normative incident investigation data about what happens in the operating room. A contrasting focus on the original concerns that guided local clinicians to initiate an inquiry into an adverse event will enable the development of measures that are the most meaningful and most likely to gain the trust of clinicians in the findings [26]. Incident investigation models that aim to develop insights about an adverse event that inform the local clinical operating system look at all aspects of human and technological interaction with the perioperative suite . Asking how local systems fit together and the nature of local constraints on perioperative care will provide a more dynamic and contextually sensitive approach to guiding incident investigation [26]. A systemic model of investigation shifts attention from what is already generally known to identifying what is knowable within the organization at the time of the event [27]. This is a hugely important distinction that is often lost on regulators.
The selection of an incident investigation model will inform how the organization chooses the types of incidents to be investigated, makes decisions about the process for engaging staff and providing feedback, and how to support the clinicians involved in the adverse event. It will also shape the type of information gathered from investigation reports. These factors are important in shaping the debriefing session format with local perioperative staff following the adverse event and its investigation. Questions should relate to a specific context where particular cues and patterns make sense and are recognizable [2, 26]. This approach will help guide future decision-making and judgments when faced with similar situations and also engender trust in present and future deliberations by management.
How to Run a Local Investigation
The decisions about the process and techniques for analyzing adverse events are best made at the local level where investigation teams reflect the workplace culture of the surgical center, its human resources, and the mix of perioperative activities [32]. Many health facilities and their surgical centers will have established structures for clinical governance and processes for the investigation of clinical adverse events. A particular method of inquiry (e.g., root cause analysis) is not prescribed here. Table 41.2 describes a commonplace approach (based on ref. [26]) that could be adapted as a triage tool for use in a range of settings to investigate different events:
Table 41.2
Triage questions and key decision points to consider when setting up an investigation that will facilitate and support the team process within the facility
Triage questions | Key decision making points |
|---|---|
What is the political landscape for inquiry? | Evaluate the existing process for the investigation of adverse events, and the track record at the facility |
Is the inquiry within the scope of your facility? | Select techniques from the available toolkit at the facility and for which there is local expertise and experience |
Who will commission a team? | Establish a core group of experienced investigators/senior clinicians to appoint a team leader and the advise team |
Who will lead the team? | Identify a senior clinician with clinical currency in the facility who is not involved in the event under review |
Who needs to be on the team? | Appoint an investigation team with the requisite knowledge of the clinical area, balancing representation across disciplines and clinical specialties, from staff not involved in the event under review |
Who is responsible and accountable for which actions during the investigation? | Determine the number of investigation team meetings required and the available time for each meeting Define the internal and external team reporting requirements, including who signs off on the final report Identify who will endorse the investigation team findings, allocate resources and support the implementation of the team recommendations across the organization Set a timeframe to evaluate the effectiveness and impact of any proposed changes to practice Organize debriefing sessions with different groups of staff at regular intervals to provide timely feedback |
Risk Assessment and Triage
Standard incident risk matrices rate an event in terms of the severity of injury to the patient and the likelihood of recurrence of the event type. There is a need for some level of triage beyond the risk assessment stage where experienced local senior clinicians not involved in the event determine what aspects of the situation warrant the most attention. The emphasis when making the decision is the knowledge of previous and current relevant challenges in the perioperative environment. In the trauma case above this was particularly relevant. The trauma physician at the facility had the foresight to convene an investigation team with the necessary senior expertise and experience to evaluate the trauma call system, the escalation process for medical review, the roles and responsibilities of the trauma team once the patient arrived in the operating room, and highlighted the importance of the trauma team leader in setting care priorities [33].
Framing the Investigation Process
The investigation needs to be flexible with the amount of time allocated tailored to the complexity of the event, recognizing if similar events have been investigated in the past. This may shorten the current inquiry. Key tasks include establishing a timeline or chronology, analyzing contributing factors to the event, and taking existing and related care delivery problems in the facility into account. It is not always possible to identify specific causal factors. The team needs to consider where the greatest benefit might be obtained in making recommendations. This was evident in the trauma case, where the trauma physician leading the investigation team had the foresight to recognize that there would be considerable benefit in having the senior clinicians on the investigation team interviewing local staff and reflecting with their peers on recent practices.
Asking Questions and Gathering Information
Beyond establishing a basic chronology of an event it is useful to identify what activities people were engaged in at the time of the event (see Table 41.3). It is essential to gather information directly from people involved in the event close to the time of the event as this will increase the opportunity to capture the immediate perceptions of what happened and what operational constraints needed to be negotiated [26]. The delivery of perioperative care is increasingly complex and contingent on the interaction between multiple team members and departments that make many adjustments to routine everyday activity and continually adapt to less than optimal conditions in order to provide safe and quality care.
Table 41.3
Asking key questions that help to analyze the constraints on normal operations at the time of the event helps to situate the actions of people in a specific and naturalistic context
Constraint | Questions to unfold everyday thinking |
|---|---|
Expectations | What was the expected outcome of the clinical intervention or activity for the patient in the perioperative environment? |
Professional standards | What were the normal parameters or standards that clinicians were expected to follow? |
Expertise and experience | What were the reasonable limits on human performance at the time? Were people working outside of their usual roles? |
Work environment | How did the people who were involved in the event identify cues and make sense of their work environment? |
Protocols and procedures | Were there any obvious adaptations of the normative care protocols that were deemed necessary at the time? |
Teamwork | Were people working independently or did the activity require some level of teamwork and cooperation? |
Attention | Where did people focus their attention and what was pointedly ignored by people in the situation? What competing demands did people need to negotiate in order to participate in the activity? |
Perception | What perception did people have of evolving changes in the immediate physical setting as the event unfolded? |
A more nuanced and tacit understanding of what the people involved were thinking at time can be obtained by asking them to retrace their actions while speaking out loud their assumptions and their perceptions of the situation as it developed [34]. This “think aloud” approach enables people to talk about things they usually would not verbalize (e.g., thoughts, feelings , reasoning, andexpectations) [35]. Thinking out loud can provide useful information about how people interpret their environment and the constraints operating in the workplace at a specific point in time. Moreover, it situates events back into the messy flow of workplace activity [6].
Facilitating Team Meetings
The investigation team will need to consider how information will be shared in face-to-face meetings as well as online in a secure manner. Clinician demands need to be weighed carefully when determining where and how often the teams need to meet and for how long. The team meeting ideally will have a facilitator of the investigation process and a senior clinical team leader to guide the clinical conversation. The individual team members each bring their own set of experiences and levels of expertise to the investigation. Rather than the team engage in a retrospective flow charting process that is prone to hindsight bias due to knowledge of the outcome of the event, it is more productive for the investigation team to put the available information from people involved in the adverse event back into the context of the unfolding situation as it was experienced [6, 8, 26]. This approach directs the inquiry toward capturing the complexities and uncertainties of why actions made sense at the time of an event. Making assumptions based on a standardized checklist of trigger questions [9, 21] runs the risk of not allowing the team to capture the nuanced perceptions of people and the variety of valid perspectives that can be derived directly from contextual information about the unfolding unexpected situation [14].
Notably, during the trauma case , the senior clinicians on the team considered that the task of categorizing the relevant factors was the remit of the patient safety manager facilitating the process. The majority of the team conversation was dedicated to a detailed analysis of local systems and the development of insights based on comparison with the team’s broad experience with similar problematic trauma presentations.
Identifying Contributing Factors
Consulting senior management at the facility early and often in the investigation process and developing a formative picture of what type of practical recommendations could realistically be implemented as an outcome of the investigation increases the likelihood that recommended changes will be taken seriously and implemented. Talking with management also reduces the risk that an adverse event might be investigated in isolation from other safety improvement activity in the department or across the hospital. If controls and corrective actions were put in place for a similar event, this is vital information for the investigation team. Arbitrary systems for deciding whether a risk is to be mitigated or removed are too disconnected from the complex and continually changing nature of the perioperative clinical setting. It is a false and dangerous assumption that risks in health care can be removed or errors completely prevented [29]. The nature of working in human systems is such that this level of predictability does not exist in a reliable form.
Recommendations that result from an incident investigation must be tested and trialed in the clinical setting [36]. This can be via formative feedback from the frontline clinicians or through simulation prior to implementation, depending on the level of complexity of the activity [37]. Simulation is an incredibly useful and visual form of event analysis. Whether using desktop, task trainer or a high fidelity surgery simulator, it can highlight the breakdowns in human performance and errors in the use of technology during the event [38, 39]. The simulation helps ensure all members of the investigation team as well as management understands what actually occurred during the event and how the team performed [40].
The Investigation Report
The team report describes the process and outcomes of the event, contributing factors, recommendations, and strategies for implementation, with timeframes for review and evaluation. It should be acknowledged that the team’s view is a limited perspective based on the available information at the time of the investigation [41]. A meeting to debrief and discuss the team’s findings with local staff across discipline and department boundaries is the single most important step. In the example of the trauma case above, there was significant email conversation between clinicians about drafting recommendations outside of the scheduled team meetings. The investigation reporting process became a vehicle for the articulation of patterns and the identification of potential solutions to the issues raised by the discussion of the case. Constructing the investigation report provided the team with a medium for inter-professional dialogue and debate that did not previously exist in the perioperative culture of the facility.
Staff Debriefings
Translating investigation reports into meaningful actions is a challenging task. In fact, in our 35 years of combined experience in being part of over 400 adverse event investigations, the investigation process is largely disconnected from everyday clinical practice and thus imposes a huge administrative burden on individuals who have ongoing operational responsibility as well as investigating the process failures that led to the adverse event. Feedback following an investigation and the implementation of strategies to implement change is not well managed [5]. Providing ongoing feedback to staff in a completely transparent manner with an interest in the event at strategic points during the investigation and debriefing after the completion of the investigation is essential if the analysis is to penetrate the local clinical workplace culture and lead to entrusting future communications [15, 17, 42]. It is additionally important to evaluate the process followed by the investigation team and to measure how effective the recommendations made by the investigation team were in addressing the challenges related to the original situation. The debriefing needs to focus on the aspects of the problematic situation that warrant the most attention in order to reduce the interference of competing agendas. An adverse event will involve many potential problems that could potentially consume large amounts of time and resource. It is useful for debriefing sessions to look beyond the event and consider the patterns and trends from similar events within the context of the facility.
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
