Wednesday, July 31, 2019

Incidence & causes of human error as a source of adverse events associated with surgical care: 56.4% adverse events were due to human error, of which cognitive error accounted for 51%

Analysis of Human Performance Deficiencies Associated With Surgical Adverse Events. James W. Suliburk et al. JAMA Netw Open. 2019;2(7):e198067, July 31, 2019, doi:10.1001/jamanetworkopen.2019.8067

Key Points
Question  What are the incidence and causes of human error as a source of adverse events associated with surgical care?

Findings  In a quality improvement study including 5365 operations, 188 adverse events were recorded. Of these, 106 adverse events (56.4%) were due to human error, of which cognitive error accounted for 99 of 192 human performance deficiencies (51.6%).

Meaning  Current systems-based approaches to improve surgical safety should be supplemented with additional focus on cognitive errors associated with surgical care.


Abstract
Importance  Potentially preventable adverse events remain a formidable cause of patient harm and health care expenditure despite advances in systems-based risk-reduction strategies.

Objective  To analyze and describe the incidence of human performance deficiencies (HPDs) during the provision of surgical care to identify opportunities to enhance patient safety.

Design, Setting, and Participants  This quality improvement study used a new taxonomy to inform the development and implementation of an HPD classifier tool to categorize HPDs into errors associated with cognitive, technical, and team dynamic functions. The HPD classifier tool was then used to concurrently analyze surgical adverse events in 3 adult hospital affiliates—a level I municipal trauma center, a quaternary care university hospital, and a US Veterans Administration hospital—from January 2, 2018, to June 30, 2018. Surgical trainees presented data describing all adverse events associated with surgical services at weekly hospital-based morbidity and mortality conferences. Adverse events and HPDs were classified in discussion with attending faculty and residents. Data were analyzed from July 9, 2018, to December 23, 2018.

Main Outcomes and Measures  The incidence and primary and secondary causes of HPDs were classified using an HPD classifier tool.

Results  A total of 188 adverse events were recorded, including 182 adverse events (96.8%) among 5365 patients who underwent surgical operations and 6 adverse events (3.2%) among patients undergoing nonoperative treatment. Among these 188 adverse events, 106 (56.4%) were associated with HPDs. Among these 106 HPD adverse events, a total of 192 HPDs (mean [SD], 1.8 [0.9] HPDs per HPD event) were identified. Human performance deficiencies were categorized as execution (98 HPDs [51.0%]), planning or problem solving (55 HPDs [28.6%]), communication (24 HPDs [12.5%]), teamwork (9 HPDs [4.7%]), and rules violation (6 HPDs [3.1%]). Human performance deficiencies most commonly presented as cognitive errors in execution of care or in case planning or problem solving (99 of 192 HPDs [51.6%]). In contrast, technical execution errors without other associated HPDs were observed in 20 of 192 HPDs (10.4%).

Conclusions and Relevance  Human performance deficiencies were identified in more than half of adverse events, most commonly associated with cognitive error in the execution of care. These data provide a framework and impetus for new quality improvement initiatives incorporating cognitive training to mitigate human error in surgery.


---
Discussion

This study found that HPDs, or human error, were identifiable in more than half of the complications occurring in major cardiothoracic, vascular, abdominal transplant, surgical oncology, acute care, or general surgical operations performed at 3 large institutions in a major academic medical center. Given that we and others report a current surgical adverse event rate of nearly 5%,12-15,17 our data suggest that more than 400 000 potentially preventable adverse events associated with HPDs occur among the nearly 17 million inpatient and ambulatory operative procedures performed in the United States annually.20
Similarities between adverse event rates in our study compared with previous studies suggest that human error remains a significant unresolved cause of adverse events in health care delivery.4-9,13-18 Specifically, compared with our currently reported adverse event rate of 3%, prior studies conducted as early as 20007,15,17,21-30 reporting a preventable or human error event rate ranging from approximately 3% to 4% support the need for interventions beyond current systems-based strategies if we are to achieve Six Sigma safety levels.
In our current analysis, cognitive error (ie, HPD class I.A.3 and class II.A-C) was the most common specific form of HPD, classified in 99 (52%) of 192 HPDs. More specifically, cognitive error was classified in 29 of 53 isolated HPD events (55%) and 70 of 139 clustered HPD events (50%). It is interesting that lack of recognition was the most prevalent cognitive error and was classified in 19% of the HPD subclassifications, potentially reflecting the paradox that the most common dangers to patient safety are those that are initially unrecognized. This paradox raises important challenges for cognitive training.
Cognitive error was likewise the most common primary cause of adverse events in clustered HPD events, being classified as such in 31 of 53 clustered HPD events (58%). Cognitive error was a potentially preventable primary cause of adverse events in 35% of technical error HPD events. In comparison, technical errors occurring in isolation were classified in only 10% of HPD events, and systems-based (class I.A.1, class I.B, and class III), communications (class IV), or teamwork (class V) HPDs; together represented only 26% of HPD events. These findings suggest the dominant role of cognitive error as a root cause of surgical adverse events, even those that would appear to be technical rather than cognitive in nature.
Considering the relative frequency of HPD types found in our study, the prevalence of cognitive error as the most frequent isolated HPD associated with adverse events may reflect the potency of cognitive error in being able to overwhelm other potential barriers to adverse events, consistent with the Swiss cheese model of the multifactorial causality of human error by Reason.31 In comparison, the frequency with which cognitive error was associated with other HPDs (100 of 192 HPDs [52%]) may also reflect the snowballing effect of HPDs, as suggested by Mold and Stein32 and Woolf et al.33
Two specific examples from our HPD analysis further highlight the potentially unrecognized role that cognitive error can also play in surgical adverse outcomes. In one case, we found that a significant technical error in a surgeon’s performance of an operation was likely precipitated by the initially unappreciated influence of the surgeon’s distraction by an outside telephone call in the operating room. The identification of distraction (ie, class II.B, lack of attention) converts an otherwise challenging technical training goal into an opportunity for behavioral training to reset following intraoperative distractions.
In a second case, a stylus that was inadvertently retained postoperatively was clearly visible but repeatedly unrecognized by radiologists in their reports prior to the patient developing a life-threatening adverse reaction. Standard event analyses would likely ascribe this outcome to clinician inattention, but we were able to further resolve this to confirmation bias (class I.A.3): the clinician dismissed their own concerns because they were not validated in official radiology reports. Like the first example, this HPD subanalysis would allow a cognitive training opportunity to teach clinicians to avoid losing their situational awareness to the convenience of alternative data.
Systems-based approaches (eg, no telephone calls in the operating room, catheter placement checklist) represent standard remedies to these HPD scenarios. However, the effectiveness of these strategies is becoming impaired by a growing checklist burnout syndrome.12,14,21-24,34 The option of cognitive training for health care practitioners as practiced in the aviation industry19 also fits with Swiss cheese model by Reason31 of multilayered human error safeguards.10,32,35,36 As noted by Gruen et al, “protocols alone are insufficient to consistently change behavior.”15

No comments:

Post a Comment