Presentation
PS11 - Redefining Human Error in Healthcare
SessionPoster Session 2
DescriptionRecent high-profile cases1,2 have drawn attention to healthcare practitioners and the role of human error in negative outcomes. While it could be assumed that healthcare professionals share a common understanding of human error, this assumption may not be correct given historical and professional differences in training received by medical subspecialties and those that manage the healthcare organization. The definition of error is important as it constrains interpretations of causes, attribution of blame, the role of chance, and the design of interventions. Thus, we believe a review of the definition of human error in healthcare and a comparison with the definition of error in other socially critical domains is warranted. This comparison seeks to identify unique perspectives from other domains regarding their definitions of human error from the literature, codifying their associated traits, and identifying recommendations for a more robust conceptualization of error for healthcare. Through this effort, we seek to facilitate conversations regarding human error in healthcare and enable individuals and organizations to reach a common understanding of human error.
We compared definitions of error from medicine with definitions from other domains including aviation, surface transportation, manufacturing, maritime, nuclear power, and space flight then identified novel aspects of these definitions. Additionally, forward and back treeing searches were utilized to identify the original sources for the definitions being presented and their domains of origin. Any definitions that could not be directly attributed to one of the previously discussed domains were categorized as general. Across the eight domains ultimately examined, thirty-seven definitions were collected with novel characteristics being codified. These characteristics (also referred to as definition items), ranged widely with differences in emphasis (e.g., people, tasks, environment, etc.) to the nature of human error itself (e.g., not result of chance, unplanned, recognizable afterward, etc.). Twenty-two unique definition items were identified and analyses were performed to assess their utilization across and within the domains in question. Two authors completed the identification process with an average agreement level of 86% (κ = .49) indicating moderate agreement. These definition items and the associated rationale proposed for their utilization within the domains are explored.
The definition of error across the domains differed greatly; with the most utilized definition item being “Related to People” with approximately 33 instances, or 89% utilization across all domains. This item reflects that a given definition of human error acknowledges human operators as having any role in influencing error without presuming fault, but instead that they are one component of the system which produces an outcome. Beyond this, the use of definition items ranged dramatically with a mean of 22.17% and a minimum of 4.05%. Definition items that fell near this mean utilization included “Related to Tasks” (17.57%), “Related to Tools” (18.92%), and “Inappropriate Outcome” (27.03%) whereas definition items such as “Not Result of Chance” (4.05%) and “Trade-Offs/ Competing Goals” (4.05%) fell at the bottom of the range. This diversity in utilization for specific definition items across the domains is indicative of unique perspectives and insights to be gained from these respective bodies of work that are potentially relevant to healthcare. Excluding the definition items identified within the Medicine domain, six of the twenty-two definition items were not found in healthcare literature. Non-utilized definition items included, “Preventable” which reflects that human error can be avoided or mitigated (i.e., it is not unavoidable or inevitable) and “Not Result of Chance” which reflects that human errors are not the result of happenstance or "acts of god" with no other influencing factors. There were nine remaining definition items which were underutilized (25%) including, “Event Chain”, “Exceeds System Tolerances”, and “Trade-Offs / Competing Goals”. Below, key perspectives of note that can be translated into healthcare are described.
The need to consider the influence of event chains.
Human error does not occur in a vacuum; each seemingly inconsequential and ordinary event that takes place has the potential to influence the system and lead to errors. Even in instances where hindsight indicates an individual could have done something differently, attributing these outcomes as being the direct fault of a given human practitioner is fundamentally naive and doesn’t advance the safety of the system as a whole. Considered within the context of surface transport, an individual’s foremost goal is typically to reach their destination in as safe and efficient a manner as possible. Despite this, they may encounter any number of hazards along their way outside their sphere of influence which may lead them to an accident. If this did occur, it would not be logical to suggest that they avoided driving in the first place or that they ought to have taken a different route to reach their destination.
The importance of well-defined, developed, and maintained compensatory mechanisms within systems.
Systems can be poised for catastrophe without mechanisms and processes in place to protect the practitioners and patients. Despite healthcare practitioners' endeavors to fulfill their roles to the best of their abilities, given their capabilities and knowledge at the time, a patient can still experience a poor and seemingly preventable outcome. While some may seek to blame healthcare practitioners or other individuals who play active roles in healthcare, this demonstrates the fallacy of many individuals’ approach to human error and their failure to account for system considerations. With clearly defined systems, including limitations and common problem points, practitioners can attempt to avoid and better prepare for complications presented; while, administrators and other organizational stakeholders work to implement the necessary compensatory mechanisms. Within nuclear power, this could be compared to a reactor not having overrides in place to power down a reaction should it reach a state of criticality. Similarly, if said overrides were experiencing an error or were otherwise in need of maintenance, then technicians could act much more conservatively during their operations with the added contextual knowledge of the system.
The influential role that competing goals play in effecting errors.
Even when accounting for event chains or presenting theoretically well-rounded systems, competing goals can still precipitate errors and undermine patient outcomes. Practitioners and other stakeholders must constantly balance goals such as maximizing patient safety, minimizing costs, and optimizing the allocation of resources while also monitoring countless other sub-tasks, goals, and processes needed to realize these aims. Each stands in opposition to the others, and when systems fail to account for these interactions, patients can and will suffer. When examined from the perspective of the manufacturing domain, this concept can be easily demonstrated. A given assembly line can only produce a set number of widgets when each of the component processes are working together efficiently. While each component process has likely been optimized to complete its task as effectively and efficiently as possible, the forward progress of one sub-goal does not undermine another.
Given the trends identified concerning definitions of human error across high-risk domains, clear gaps exist within the healthcare space. By reframing the hard-won knowledge generated in these areas to the issues presented within healthcare, steps can be taken to create a shared understanding between individuals and organizations regarding human error; and ultimately more effectively combat these issues. These can be achieved by implementing key perspectives such as the influence of event chains and cumulative interactions; the importance of well-defined, developed, and maintained compensatory mechanisms within systems; and the role that competing goals play in affecting errors.
We compared definitions of error from medicine with definitions from other domains including aviation, surface transportation, manufacturing, maritime, nuclear power, and space flight then identified novel aspects of these definitions. Additionally, forward and back treeing searches were utilized to identify the original sources for the definitions being presented and their domains of origin. Any definitions that could not be directly attributed to one of the previously discussed domains were categorized as general. Across the eight domains ultimately examined, thirty-seven definitions were collected with novel characteristics being codified. These characteristics (also referred to as definition items), ranged widely with differences in emphasis (e.g., people, tasks, environment, etc.) to the nature of human error itself (e.g., not result of chance, unplanned, recognizable afterward, etc.). Twenty-two unique definition items were identified and analyses were performed to assess their utilization across and within the domains in question. Two authors completed the identification process with an average agreement level of 86% (κ = .49) indicating moderate agreement. These definition items and the associated rationale proposed for their utilization within the domains are explored.
The definition of error across the domains differed greatly; with the most utilized definition item being “Related to People” with approximately 33 instances, or 89% utilization across all domains. This item reflects that a given definition of human error acknowledges human operators as having any role in influencing error without presuming fault, but instead that they are one component of the system which produces an outcome. Beyond this, the use of definition items ranged dramatically with a mean of 22.17% and a minimum of 4.05%. Definition items that fell near this mean utilization included “Related to Tasks” (17.57%), “Related to Tools” (18.92%), and “Inappropriate Outcome” (27.03%) whereas definition items such as “Not Result of Chance” (4.05%) and “Trade-Offs/ Competing Goals” (4.05%) fell at the bottom of the range. This diversity in utilization for specific definition items across the domains is indicative of unique perspectives and insights to be gained from these respective bodies of work that are potentially relevant to healthcare. Excluding the definition items identified within the Medicine domain, six of the twenty-two definition items were not found in healthcare literature. Non-utilized definition items included, “Preventable” which reflects that human error can be avoided or mitigated (i.e., it is not unavoidable or inevitable) and “Not Result of Chance” which reflects that human errors are not the result of happenstance or "acts of god" with no other influencing factors. There were nine remaining definition items which were underutilized (25%) including, “Event Chain”, “Exceeds System Tolerances”, and “Trade-Offs / Competing Goals”. Below, key perspectives of note that can be translated into healthcare are described.
The need to consider the influence of event chains.
Human error does not occur in a vacuum; each seemingly inconsequential and ordinary event that takes place has the potential to influence the system and lead to errors. Even in instances where hindsight indicates an individual could have done something differently, attributing these outcomes as being the direct fault of a given human practitioner is fundamentally naive and doesn’t advance the safety of the system as a whole. Considered within the context of surface transport, an individual’s foremost goal is typically to reach their destination in as safe and efficient a manner as possible. Despite this, they may encounter any number of hazards along their way outside their sphere of influence which may lead them to an accident. If this did occur, it would not be logical to suggest that they avoided driving in the first place or that they ought to have taken a different route to reach their destination.
The importance of well-defined, developed, and maintained compensatory mechanisms within systems.
Systems can be poised for catastrophe without mechanisms and processes in place to protect the practitioners and patients. Despite healthcare practitioners' endeavors to fulfill their roles to the best of their abilities, given their capabilities and knowledge at the time, a patient can still experience a poor and seemingly preventable outcome. While some may seek to blame healthcare practitioners or other individuals who play active roles in healthcare, this demonstrates the fallacy of many individuals’ approach to human error and their failure to account for system considerations. With clearly defined systems, including limitations and common problem points, practitioners can attempt to avoid and better prepare for complications presented; while, administrators and other organizational stakeholders work to implement the necessary compensatory mechanisms. Within nuclear power, this could be compared to a reactor not having overrides in place to power down a reaction should it reach a state of criticality. Similarly, if said overrides were experiencing an error or were otherwise in need of maintenance, then technicians could act much more conservatively during their operations with the added contextual knowledge of the system.
The influential role that competing goals play in effecting errors.
Even when accounting for event chains or presenting theoretically well-rounded systems, competing goals can still precipitate errors and undermine patient outcomes. Practitioners and other stakeholders must constantly balance goals such as maximizing patient safety, minimizing costs, and optimizing the allocation of resources while also monitoring countless other sub-tasks, goals, and processes needed to realize these aims. Each stands in opposition to the others, and when systems fail to account for these interactions, patients can and will suffer. When examined from the perspective of the manufacturing domain, this concept can be easily demonstrated. A given assembly line can only produce a set number of widgets when each of the component processes are working together efficiently. While each component process has likely been optimized to complete its task as effectively and efficiently as possible, the forward progress of one sub-goal does not undermine another.
Given the trends identified concerning definitions of human error across high-risk domains, clear gaps exist within the healthcare space. By reframing the hard-won knowledge generated in these areas to the issues presented within healthcare, steps can be taken to create a shared understanding between individuals and organizations regarding human error; and ultimately more effectively combat these issues. These can be achieved by implementing key perspectives such as the influence of event chains and cumulative interactions; the importance of well-defined, developed, and maintained compensatory mechanisms within systems; and the role that competing goals play in affecting errors.
Event Type
Poster Presentation
TimeTuesday, April 14:45pm - 6:15pm EDT
LocationFrontenac Foyer