Impact and prevention of errors in endovascular treatment of unruptured intracranial aneurysms

Abstract

Background

Preventing errors and complications in neurointervention is crucial, particularly in the treatment of unruptured intracranial aneurysms (UIAs), where the natural history is generally benign, and the margin of treatment benefit small. We aimed to investigate how neurointerventionalists perceive the importance and frequency of errors and the resulting complications in endovascular UIA treatment, and which steps could be taken to prevent them.

Methods

An international multidisciplinary survey was conducted among neurointerventionalists. Participants provided their demographic characteristics and neurointerventional treatment volume. They were asked about their perceptions on the importance and frequency of different errors in endovascular UIA treatment, and which solutions they thought to be most effective in preventing these errors.

Results

Two-hundred-thirty-three neurointerventionalists from 38 countries participated in the survey. Participants identified errors in technical execution as the most common source of complications in endovascular UIA treatment (40.4% thought these errors constituted a relatively or very large proportion of all complication sources), closely followed by errors in decision-making/indication (32.2%) and errors related to management of unexpected events (28.4%). Simulation training was thought to be most effective in reducing technical errors, while cognitive errors were believed to be best minimized by abandoning challenging procedures, more honest discussion of complications and better standardization of procedure steps.

Conclusion

Neurointerventionalists perceived both technical and cognitive errors to be important sources of complications in endovascular UIA treatment. Simulation training, a cultural change, higher acceptance of bail-out strategies and better standardization of procedures were perceived to be most effective in preventing these.

Introduction

Medicine is an imperfect science, and this is particularly true for neurointervention. The prevalence of manifest or imminent complications due to human errors in neurointervention has been described to be as high as 12%,¹ many of which seem predictable and preventable.

The neurointerventional literature generally tends to downplay the importance of complications. Case-series reporting the use of new devices for instance often define complications as major events that lead to permanent severe neurological deficits, and do not capture so-called “minor” events such as non-disabling and/or temporary neurological deficits. Complications that do not result in any neurological deficit (e.g. groin hematomas or femoral pseudoaneurysms) are often ignored altogether. Fear of being judged by colleagues and litigation concerns further complicate the “medical error culture”.² Nevertheless, preventing errors and complications is crucial, especially in high-stakes fields such as neurointervention, both to improve procedural safety and to reduce the mental strain for operators themselves. This is particularly true for treatment of unruptured intracranial aneurysms (UIA), where the natural history is generally good, and the margin of benefit with treatment is small – in other words, there is “little to gain and much to lose”. We aimed to determine how neurointerventionalists perceive the importance and frequency of errors and the resulting complications in endovascular UIA treatment, and to investigate which steps could be taken to prevent these.

Methods

We conducted an international web-based survey using Qualtrics.com (an ISO 27001 certified online survey tool with real-time feedback that was previously used for survey-based stroke research^3,4) to explore how neurointerventionalists approach complications in endovascular UIA treatment. In the first part of the survey, participants were asked to provide some baseline information (age, gender, specialty, career stage, years of neurointerventional experience, personal and institutional elective endovascular aneurysm treatment volume). The survey was deliberately sent to neurointerventionalists in early and advanced career stages (residents, fellows, junior and senior staff) to ensure a broad and inclusive spectrum of responses. In the second part, physicians answered specific questions about (1) their personal method of estimating and communicating complication risks and (2) the importance and frequency of different errors that can lead to complications in endovascular UIA treatment, and which solutions they thought to be most effective in preventing these errors. A closed-ended question design with available free text response option was used for all questions (see Supplementary material for a detailed printout of the original survey questions). Anonymized, personalized links were used to avoid duplicate responses and participants were required to answer all questions (“force response mode”), which ensured that there were no missing data. Survey data were summarized using descriptive statistics and compared between physicians in early career stages (residents, fellows & junior staff [within 5 years from board certification]) and senior staff (beyond 5 years from board certification). Data analysis was performed in Stata 15.1. All participants gave their consent prior to answering the survey, and participation was voluntary. Because no patient data was used, an ethics approval was not necessary for this study.

Results

Response data were obtained from April 3 to 24, 2020. Out of 1302 neurointerventionalists, 369 neurointerventionalists iniated the survey (survey initiation rate: 28.3%), 233 of which completed the survey, herein referred to as “survey participants” (completion rate: 63.1%). The final sample represented a broad and inclusive spectrum of neurointerventionalists from 38 countries (median age 47 years [IQR 41–55], 25/233 [10.7%] females). The specialty distribution of invited and participating physicians was similar; participants baseline characteristics are shown in Table 1.

Table 1.

Survey participants’ baseline characteristcs (N = 233).

Variable
Age in years – median (IQR)	47 (41–55)
Female sex	25 (10.73%)
Career stage
Resident	4 (1.72%)
Fellow	7 (3.00%)
Junior staff (within 5 years from board certification)	22 (9.44%)
Senior staff (greater than 5 years from board certification)	200 (85.84%)
Background specialty
Neuroradiology	151 (64.81%)
Neurology	15 (6.44%)
Neurosurgery	65 (27.9%)
Other	2 (0.86%)
Years of neurointerventional experience – median (IQR)	15 (10 – 21)
Personal annual endovascular UIA treatment volume
<10	25 (10.73%)
10–20	51 (21.89%)
20–30	57 (24.46%)
30–40	29 (12.45%)
40–50	17 (7.30%)
>50	54 (23.18%)
Institutional annual UIA treatment volume
<20	25 (10.73%)
20–40	40 (17.17%)
40–60	52 (22.32%)
60–80	33 (14.16%)
80–100	25 (10.73%)
>100	58 (24.89%)
Frequency of UIA case discussion in a multidisciplinary board
Not at all	14 (6.0%)
Only selected UIA cases	87 (37.3%)
All UIA cases	132 (56.7%)

IQR: interquartile range; UIA: unruptured intracranial aneurysm.

Neurointerventionalists’ approach to estimating and communicating complication risks in endovascular UIA treatment

Many survey participants (72/233 [30.9%]) stated that they rely on their personal treatment complication rate when estimating the risk of treatment complications in a given patient with UIA. An almost equally high number of participants (67/233 [28.8%]) stated that they use a combination of personal, institutional and published complication rates to estimate the risk of treatment complications for an individual case. Of the remaining physicians, 54/233 (23.2%) used their hospital-specific complication rate and 40/233 (17.2%) relied on published complication rates from the literature to estimate the risk of treatment complications in an individual patient. When asked about their method of communicating complication risks prior to endovascular UIA treatment to patients, the vast majority of neurointerventionalists (206/233 [88.4%]) stated that rather than telling their patients a “default” complication rate, they modify the complication estimate based on individual patient factors. Most physicians (211/233 [90.6%]) said that they tell their patients a concrete number when they communicate complication risks rather than using more general terms such as “unlikely”, “possible” or “probable”.

Perceived frequency of different errors leading to complications in endovascular UIA treatment and possible solutions to prevent them

Participants were confronted with seven specific types of errors (Table 2), and stated how frequent they thought these errors lead to complications in endovascular UIA treatment. Errors in technical execution were hereby identified as the most common source of complications, followed by errors related to management of unexpected events and errors in decision-making/indication (Figure 1). Physicians in advanced career stages generally thought most types of errors to be more common compared to physicians in earlier career stages. For instance, only 6.1% physicians in early career stages (residents, fellows and junior staff) thought that errors in decision making formed a relatively or very large proportion of all errors that lead to complications in endovascular UIA treatment, as opposed to 9.5% of the more experienced physicians (senior staff). Similar patterns were observed for errors related to management of unexpected events (12.1% vs. 31.0%), errors due to lacking willingness to accept imperfect results (36.4% vs. 37.5%), antiplatelet management errors (12.1% vs. 26.0%), technical errors (27.2% vs. 32.5%), errors in device choice (12.1% vs. 21.5%), and errors in decision making/indication (24.3% vs. 33.5%). For stratified answers by physician career stage see Supplementary Figures 1 and 2. In a next step, physicians were asked what they considered the most effective measures to prevent these seven types of errors. Answers were chosen from a pre-defined set of response options and are summarized in Figures 2 and 3. Most of the solutions physicians perceived as most effective were related to better physician education and a “cultural change” (e.g. more multidisciplinary rounds to discuss UIA cases in order to avoid errors in decision-making (Figure 2(a)), better education on the limitations of devices and less pressure from peers to use new devices in order to avoid errors in device choice (Figure 2(b)), greater willingness to abandon challenging procedures earlier (Figure 2(c)) and more honest presentation and discussion of complications in meetings to avoid errors in the management of unexpected events (Figure 3(c)). Specific simulation training in individual patient anatomy was considered to be highly effective in the prevention of technical errors (Figure 2(c)). Other commonly chosen answers referred to better standardization of procedure steps (e.g. better guidelines on which patients to treat to avoid errors in decision-making (Figure 2(a)), standardized protocols for antiplatelet management (Figure 3(a)) and standard operating procedures for rare, unexpected events (Figure 3(c))). Relatively few physicians thought that improvement of devices or development of better devices would offer increased effectiveness in reducing errors that can cause complications in endovascular UIA treatment. Responses of physicians in early career stages (residents, fellows, junior staff) and advanced career stages (senior staff) were hereby very similar for most errors (Supplementary Table 1).

Table 2:

Different types of errors that can potentially lead to complications in endovascular UIA treatment.

Type of error	Examplea
Errors in decision-making/indication	“This aneurysm should not have been treated in the first place”
Errors in device choice	“This aneurysm would have been better treated with stent-assisted coiling rather than flow diversion”
Errors in technical execution	Occlusion of the parent vessel due to prolapse of an undersized coil
Errors in antiplatelet management	Suboptimal choice of antiplatelet agents, dosing, timing
Errors due to lacking willingness to accept imperfect results	The coiling looks good but you attempt to “make it perfect” and put another coil and then end up dislodging coil in the parent vessel
Errors related to management of unexpected events	A thrombus is forming in the stent and you fail to react timely and appropriately with intensified antiplatelet therapy
Errors in long-term management	Loss of follow up, antiplatelets stopped to early, recurrence on follow-up imaging not recognized

^aExamples as they were used in the survey questionnaire.

Figure 1.

Perceived frequency of specific types of errors leading to complications in endovascular UIA treatment.

Figure 2.

Perceived possible solutions to prevent errors leading to complications in endovascular UIA treatment. Participants were asked to choose 3 of the pre-specified answer options in (a and b), and 2 options in (c). Participants’ most common top choices are highlighted in light blue.

Figure 3.

Perceived possible solutions to prevent errors leading to complications in endovascular UIA treatment (continued). Participants were asked to choose 3 of the pre-specified answerer options in (c), 2 options in (a), and 1 option in (b) and (d) respectively. Participants’ most common top choices are highlighted in light blue.

Discussion

Neurointerventionalists in this study thought that errors in decision-making and technical execution, and the desire to pursue “perfect” treatment results were the most common errors leading to complications in endovascular UIA treatment. Solutions that were perceived as effective in reducing these errors were mostly related to improvements in education, simulation training, changes in the conceptual approach to cases, and better standardization of procedure steps.

Broadly speaking, errors in neurointervention can be system-related (e.g. understaffed teams, insufficiently maintained equipment) or operator-related. The latter ones can be further divided into technical errors (i.e. errors in technical execution) and those due to non-technical skills (“cognitive errors”). Errors may or may not result in complications, but even if they don’t (“near misses”), they constitute valuable learning opportunities and should be analyzed as thoroughly as errors that resulted in a complication. On the other hand, not every complication is the result of an operator or system error; some complications are due to random chance; they occur even in well-executed procedures and are thus unpreventable. It is therefore not possible to prevent all complications. The focus of this study was preventable complications, or more precisely, operator- and system-related errors that can result in preventable complications. More than 85% of the participants were senior staff, possibly because experienced physicians are more comfortable facing and discussing complications than junior physicians, and this might have biased our results. Compared to more junior physicians, senior staff also thought most error types to lead more commonly to complications than junior staff, potentially suggesting that less experienced physicians might tend to underestimate the frequency and impact of different types of errors in endovascular UIA treatment. While technical errors were considered a frequent cause of complications in this study, cognitive mishaps, namely errors in decision-making and those related to lacking willingness to accept imperfect results, were considered equally important. Reducing technical errors through improvement of technical skills is relatively easy and can be reliably achieved through training in flow models or a simulation environment.^5–7 Simulation training can also help avoid errors during extreme situations such as the COVID-19 pandemic, as it allows the neurointerventional team to practice relevant workflows, e.g. patient transport through contaminated and “clean” areas or donning and doffing personal protection equipment.⁸ Thus, it would be desirable to make simulation training more widely available to physicians. Ideally, simulators would be part of the standard equipment in interventional centers. Surgical specialties have recognized the value of simulation training early on, and simulation is now integral part of the surgical training curriculum,^9–11 but access for neurointerventionalists with radiology background, who constituted the majority of the respondents, to such tools might be limited. Preventing cognitive errors, which were more commonly thought to lead to complications, is much harder but not impossible. Routine consideration of alternatives and cognitive “stopping points” for instance are effective strategies that can help to minimize cognitive errors.^12,13 The strategies neurointerventionalists in this survey considered most effective were mostly related to a “change in error culture”, such as more open and honest discussions of errors and complications with peers, or a higher acceptance of angiographically “imperfect” results. The second most common theme was standardization and better guidelines for certain procedural situations, which could help to reduce the cognitive load on operators and thereby allow them to direct their attention to other tasks. Both a culture change and a higher degree of standardization can in theory be achieved without any financial resources or infrastructural changes, but they require a substantial change in the mindset of most neurointerventionalists, which might not be always achievable. Surprisingly, system factors, e.g. better equipment and/or access to devices was considered to contribute very little to UIA treatment complications. Errors in technical execution were thought to be most effectively prevented by simulation training, but interestingly, a higher willingness to abandon challenging procedures was considered almost equally important, suggesting that cognitive factors might play an equally important role in technical errors as manual skills.

Limitations

This study has several limitations that deserve comment. First, participants of this survey were selected based on personal networks and academic collaborations, which might not be representative of the neurointerventional community as a whole, although care was taken to approach a broad and inclusive international audience. Second, we provided a pre-defined set of questions and answer options, which might not fully capture the complex nature of cognitive and technical errors, the resulting complications and the causal relationships between the two. Third, since we only asked for subjective frequency estimates (e.g. very small/relatively small proportion of all errors) for the prevalence of different error types but not for exact percentage estimates, we could not compare physicians’ subjective estimates to published complication rates in the literature. However, such a comparison is difficult to begin with, since most publications focus on device-specific and/or procedure specific complication rates rather than “conceptualizing” different types of complications. Fourth, this study investigated the perception of errors and complications in endovascular UIA treatment, which is an elective procedure, and the responses might have been different had we focused on another, more time-critical task such as acute ischemic stroke treatment. Clearly, our results are subjective in nature and should therefore be interpreted with caution. Nevertheless, they can potentially provide valuable starting points for neurointerventionalists, professional societies and industry partners who aim to improve the safety profile of endovascular UIA treatment.

Conclusion

Neurointerventionalists perceived both technical and cognitive errors to be important sources of complications in endovascular UIA treatment. Simulation training was considered most effective in reducing technical errors, while a more open discussion of complications, higher acceptance of bail-out strategies and better standardization of procedure steps were thought to best prevent cognitive errors.

Declaration of conflicting interests

The author(s) declared the following potential conflicts of interest with respect to the research, authorship, and/or publication of this article: Johanna Ospel is supported by the University of Basel Research Foundation, Julia Bangerter Rhyner Foundation and Freiwillige Akademische Gesellschaft Basel. Mayank Goyal is a consultant for Medtronic, Stryker, Mentice. The remaining authors have nothing to disclose.

Ethical approval

Since this study did not involve and human patients or animals, an ethics approval was not required. All survey respondents gave their informed consent and agreed to participate in the study.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

ORCID iDs

Johanna Maria Ospel https://orcid.org/0000-0003-0029-6764

Anil Gopinathan https://orcid.org/0000-0002-8192-1287

Supplemental material

Supplemental material for this article is available online.