Although Leape1 called attention to the need for a reduction in medical error over a decade ago, in 2000, the Institute of Medicine estimated the annual cost of medical error in the U.S. health care system at over 9 billion dollars and more than 100,000 patient lives.2 Far from being immune to medical error, the emergency department is one of the top three hospital areas prone to medical error.3 Approximately 18 million errors and 360,000 adverse events occur annually in U.S. emergency departments.4
The emergency department is a growing and important area for clinical education. Emergency medicine now supports more than 130 accredited residency programs with over 1,250 entry-level emergency medicine residency positions.5 The emergency department is also the educational host to medical students and residents from many other clinical disciplines. In most academic emergency departments, just as in other areas of teaching hospitals, residents are the primary interface between the system and the patient and provide a substantial portion of direct patient care.
Studies of nonemergency medicine training programs provide strong evidence that residents err in every clinical context.6–8 Residents are particularly vulnerable to the effect of errors on their clinical practice.8 Maladaptive behaviors may result when errors occur during residency if the proper faculty oversight, professional modeling, and productive discussion are absent. Once ingrained, these maladaptive practice patterns may become pervasive and lifelong barriers to quality practice. Therefore, it is imperative that a residency training experience encourages the resident to develop positive error management strategies and responses via constructive behavioral change. To create such a residency experience, educators must understand how residents respond to their medical errors, and how elements of the residency program and environment that are amenable to intervention influence residents' responses to error. In this study we sought to determine residents' responses to medical errors in terms of emotional state, learning response, and disclosure; and to examine associations between residents' responses to medical error and perceptions of their training programs.
A resident's response to medical error is moderated by the resident's personal characteristics, the nature of the error, and the context in which the resident works. While each resident's personal characteristics, beliefs, attitudes, experience, and culture influence his or her response to medical error, we do not discuss these elements in this article. The nature of the error influences the resident's response to the error just as the type of error influences how the lay public feels the error should be handled.
The context in which the resident works also mediates the resident's response to the medical error. The attitude of the training program toward resident error, role-modeling by faculty, institutional response to error, and the institution's cultural milieu dictating normative behavior in response to medical error may all influence the resident's response to medical error.9
Part of the response to the error is attribution of the cause of the error. The error may be attributed to some aspect of oneself or to something one has control over (intrinsic attribution). Alternatively, the error may be attributed to something external to oneself over which one does not believe one has control (extrinsic attribution).
Resident error-based learning is a multidimensional construct encompassing both constructive and defensive behavior change.8 Constructive behavior change promotes improvement and/or development of positive practice patterns to prevent future error. Defensive behavior change includes behaviors to protect the resident from reproach or attribution to avoid negative consequences of error. Constructive and defensive behavior changes are dimensions of error-based learning rather than endpoints of a single continuum, so it is possible for a resident to respond to an error with both constructive and defensive behavior change. For example, if the resident mistakenly misreads an ankle X-ray, the resident may study ankle X-rays to improve reading ability (constructive) and blame the radiologist for not reading the X-ray more quickly (defensive).
Study location, participants, and data collection
To determine how emergency medicine residents respond to medical error, in 2003 we conducted a survey of 54 residents at two accredited emergency medicine residency programs in the eastern United States. Two of us (AH and RR) administered a paper-and-pencil questionnaire to all emergency medicine residents during conferences in the two training programs. We asked residents after completing the questionnaires to put them in an anonymous envelope and seal it. We used a standard worksheet software to enter all the results from both institutions with no program or individual identifiers. Free text answers were transcribed verbatim. The Institutional Review Board at both institutions approved this study
Based on a literature review,1,2,8,10,11 faculty review, and pilot testing, we developed a questionnaire intended to assess residents' responses to errors and residents' characteristics. Questions designed to assess the extent of constructive and defensive responses to error were developed based on previous literature.8 Based on our conceptual model, we designed questions to assess three domains of response to error: emotional response, behavioral change, and error disclosure. We further identified three distinct types of behavioral change: information seeking, vigilance, and defensive practices. Emergency medicine faculty at the two institutions reviewed the items and we modified the questionnaire accordingly. Next, we pilot tested the questionnaire on a 20% sample (ten residents) at the University of North Carolina at Chapel Hill and modified the questionnaire to improve question clarity, comprehension, and response ranges.
To encourage residents to recall a nontrivial error, we asked them to provide a free-text description of their most significant medical error using the Institute of Medicine's definition of medical error as a prompt:
the failure of a planned action to be completed as intended (i.e., error or execution), as the use of a wrong plan to achieve an aim (i.e., error of planning), or as an act or omission for which the resident felt responsible that had serious or potentially serious consequences for the patient and that would have been judged wrong by knowledgeable peers at the time it occurred.2, p.28
To sharpen the residents' memory of the error, we prompted them to categorize the types, causes, and severity of the error. Next, we used five-point Likert scales (1 = Strongly disagree, 5 = Strongly agree) and closed questions to assess residents' responses to the errors they described: emotional response, behavioral change, and error disclosure. Finally, we asked residents to report basic characteristics such as age, gender, and year of training.
To better understand the associations between institutional, contextual, and error characteristics with the residents' responses to the errors, we created scales based on the domains from which the questions were originally created (emotional response, behavioral change, and error disclosure). For each of the domains, the questions from that domain were entered into a scale. To create reliable scales, we removed questions with the lowest interitem covariance and the most negative influence on the scale's Cronbach's alpha from the scale until only questions that contributed to the scale's reliability remained.
We performed univariate and bivariate descriptive statistics using chi-square and t-tests. To determine the associations between the resident responses to error described by the scales and the other domains in our model, we examined bivariate associations using the nonparametric Somers' D. We were unable to adjust for autocorrelation within program because we sampled only two residency programs. Although the institutions did differ along three dimensions (identification that the error was caused by job overload; resident responded to error by increasing communication; and resident had to disclose the error to the legal department), we found no differences in the associations examined in this study by institution. We performed all analyses using Stata 8.0/SE (Stata Corp., College Station, Texas).
Forty-three of the 54 eligible residents (80%) returned questionnaires. Of these respondents, one returned a blank questionnaire and two identified no errors in their practice and gave only demographic information. (See Table 1 for respondents' characteristics.) We analyzed responses from the 40 respondents who documented an error and responded to the questionnaire.
Description of errors and outcomes
Residents reported that 25 (63%) errors occurred while they were working in the emergency department, but errors occurred in other hospital environments as well: eight (20%) in the intensive care unit and seven (18%) on inpatient wards. Twenty-six (65%) errors took place during the intern year. Fully 25 (63%) errors occurred with patients between the ages of 17 and 60 years, while 13 (33%) errors occurred with patients older than age 60 years. Residents self-classified their errors into the following categories: 13 medication errors (33%), 11 diagnostic errors (28%), eight errors in evaluation and treatment (20%), four procedural errors (10%), and three communication errors (8%). Residents also cataloged the errors into nonexclusive categories. The three most frequent error types reported were 20 treatment administrations (50%), 19 failures of communication (48%), and 16 delays in diagnosis (40%). Because of the errors, nine (23%) patients deteriorated clinically and five (11%) patients suffered a severe disability. Seven (18%) patients died as a result of the error. In addition to clinical deterioration, 13 (33%) patients suffered physical discomfort, ten (25%) suffered emotional distress, 11 (28%) required extra therapy, six (15%) required extra monitoring, nine (23%) required extra procedures, and seven (18%) had an increased length of stay.
Causes of errors
Residents had substantial intrinsic and extrinsic attribution of errors (see Table 2). In terms of intrinsic attribution,22 (55%) residents identified inadequate experience, 20 (51%) cited inadequate knowledge, and 16 (40%) felt they had missed the warning signs. With respect to extrinsic attribution, 17 (43%) identified having other things to take care of, ten (25%) cited the complexity of the case itself as causing the error, and eight (20%) noted that the presentation was atypical.
Residents' responses to error
Most residents experienced some negative emotions as a result of the error: 27 (68%) experienced remorse, 21 (53%) guilt, 23 (58%) inadequacy, and 22 (55%) frustration. (see Table 3)
Residents reported that their errors resulted in significant changes in their learning behaviors. A moderate number of residents reported they increased their information seeking behaviors: 11 read more (28%), 12 (30%) sought advice from peers, 14 (35%) sought advice from senior staff, and 13 (33%) asked supervisors (see Table 3). Increased vigilance was a more common behavioral response: 32 (81%) reported increased attention to details and 20 (50%) personal confirmation of data. Although residents did not engage in many positive changes in practice, residents did not engage in negative changes in practice either: only five ordered more test more tests (13%), two (5%) kept errors to themselves, none started seeing fewer patients, and two (5%) began avoiding similar patients.
Only 33 residents (83%) reported disclosing the error to another party. Residents informed the responsible attending in 27 (71%) cases and discussed the error with other physicians in nine (23%) cases (see Table 3). In 17 cases (47%), the resident discussed the error with another physician who was not involved with the case. In 19 (53%) cases, residents informed a friend or spouse of the error. Surprisingly, residents disclosed the error to patients and families less than they disclosed the error to any of the other groups—in only ten (28%) cases.
Fully 35 (88%) residents felt they were moderately to completely responsible for the error. Of these, 16 (41%) felt they were completely responsible for the errors. As part of their acceptance of responsibility for the error, 36 (90%) residents were self-critical of their performances and promised to do things differently next time.
Residents who reported that the error was more attributable to a lack of experience reported a 26% (95% CI, 5–47%) more negative emotional response to the error (see Table 4). A negative emotional response was also 15% (95% CI, −8–38%) associated with job overload, 23% (95% CI, 5–41%) associated with a perceived lack of an institutional support and 21% (95% CI, 3–40%) associated with discussion of the error at Morbidity and Mortality conference (M &M).
Residency is a time of intense learning and establishment of lifelong learning patterns and best practices. It is also a time of progressive responsibility in clinical decision making and professional development.12 Understanding the effect of mistakes on residents' behavior is critical to helping residents develop successful error management techniques.
Our results suggest that errors that occur during clinical training change learning behavior and have a substantial emotional impact. Negative emotions such as anger, guilt, remorse, emotional distress, frustration and inadequacy were common responses. These negative emotions were also associated with identifiable institutional (unsupportive institutional environment), training (lack of experience), and educational factors (presentation at M & M). Each of these factors deserves attention and directed effort by faculty and administration to mitigate the associated negative emotions.
Our study highlights that M & M is one particularly important area for potential improvement. Although M & M is the most common error-related conference in emergency medicine,10 its format varies widely both among specialties13 and within emergency medicine.10 Recent data suggest that professional modeling of error acknowledgement and explicit discussion of errors are rare.13 Consistent with the findings of Orlander et al.,14 our exploratory data suggest that M &M elicits negative emotional responses. Directed efforts by the community of emergency medicine educators may substantially improve the educational focus of the conference and mitigate negative emotional responses.
Another key area for directed improvement efforts are the error types that result in defensive practice changes associated with two major groups of errors, those attributed to systems of care and to personal practices. Errors attributable to systems are the focus of broadly based institutional reforms endorsed by national organizations such as Leapfrog,15 the National Quality Forum,16 and the National Patient Safety Foundation.17 We know from our results and those previously published in internal medicine8 that erring residents' behaviors became more defensive when they perceived the institutional environment as unsupportive. What we do not know is what specifically predicts defensive practice changes. Of particular interest to educators are the personal practice errors: those related to lack of experience, clinical judgment, evaluation and treatment, and diagnosis. Understanding how these types of errors predispose residents toward defensive practice changes gives faculty a powerful tool to mitigate negative learning behaviors.
Emergency medicine educators recognize that the quality and safety of the care delivered by clinicians depend substantially on the performance capability of the organizational systems in which they work.18 While individual clinician's competence remains important, critical elements in providing high-quality care include organizational systems' abilities to prevent errors; coordinate care among settings and practitioners; and ensure that relevant, accurate information is available when needed.2 Although empirical research provides little guidance for building “error-free” clinical care systems,19 theory and research suggest that leadership and culture play critical roles in shaping the informal norms that foster open dialogue, continuous improvement, and organizational learning.20–23 More research is needed that examines the organizational systems and processes that promote residents' constructive learning behavior and mitigate defensive or counterproductive learning behavior.
Our study had several limitations: a biased estimate of error types, self-reported perceptions of cause and context, a small sample from only two institutions, and a cross-sectional design preventing determination of causality. We asked residents to describe their most significant medical error. This approach was selected to encourage the resident to identify an emotionally meaningful error to which the resident was most likely to have a response. An alternate strategy such as asking for the most recent error would most likely have skewed the data toward low-severity errors, based on error frequency,4 and risked having many errors that residents felt were too trivial to warrant behavioral change or emotional response. Although we are ultimately interested in assessing how the context and the cause of the error influence the resident's response, our study assessed only the resident's perception of the cause and the context and may not be a true reflection of the cause or the context. Indeed, altering the context in which the resident learns or the systems designed to eliminate causes of errors may do little to alter the resident's perception. Nonetheless, the goal of any future intervention will be to alter the resident's response to error, so the resident's perception is the most important measure. Our small sample size and limited diversity of training environments reduced the variation of context and limited our ability to make inferences about the effects of context. Studies into more diverse contexts with greater diversity of residents are needed to further identify the extent and the limitations of the contextual influence on residents' responses to medical error. Finally, the cross-sectional study design did not allow us to make inferences about causality. The lack of current understanding, however, urges a thorough exploration of associations so that empirically informed interventions can be designed and tested.
The emergency department has been implicated as an area where the highest percentage of negligent errors occur,3 as well as an environment whose unique operating characteristics or error producing conditions make errors more likely.24 Residents are particularly vulnerable to error-producing conditions and to developing maladaptive behaviors due to error. Currently, residents often respond to medical errors inappropriately—such as not disclosing the error to the attending. Improving residents' responses to error requires understanding the effects of the training context on residents' responses. Some long-held approaches, such as M & M, may need to be restructured to be beneficial. Errors occur even after residency and resident training must include not only how to do things correctly, but how to respond constructively when things are done incorrectly. This often neglected component of residents' education demands a rigorous understanding of the effects of errors on residents' behavior. Resident educators need to examine the influence that their learning environment has on how residents process and respond to medical errors.
1Leape LL. Error in medicine. JAMA. 1994;272:1851–57.
2Kohn LT, Corrigan J, Donaldson MS, McKenzie D. To Err is Human: Building a Safer Health System. Washington, DC: National Academy Press; 2000:xxi, 287.
3Thomas EJ, Studdert DM, Burstin HR, et al. Incidence and types of adverse events and negligent care in Utah and Colorado. Med Care. 2000;38:261–71.
4Fordyce J, Blank FS, Pekow P, et al. Errors in a busy emergency department. Ann Emerg Med. 2003;42:324–33.
5Hamilton G. Professing on a milestone in emergency medicine. SAEM Newsl 2003;XV(6):5.
6Shaughnessy AF, Nickel RO. Prescription-writing patterns and errors in a family medicine residency program. J Fam Pract. 1989;29:290–5.
7Gratton MC, Salomone JA, 3rd, Watson WA. Clinically significant radiograph misinterpretations at an emergency medicine residency program. Ann Emerg Med. 1990;19:497–502.
8Wu AW, Folkman S, McPhee SJ, Lo B. Do house officers learn from their mistakes? JAMA. 1991;265:2089–94.
9Bosk C. Forgive and Remember: Managing Medical Failure. Chicago: University of Chicago Press, 1979.
10Hobgood CD, Ma OJ, Swart GL. Emergency medicine resident errors: identification and educational utilization. Acad Emerg Med. 2000;7:1317–20.
11Hobgood C, Peck CR, Gilbert B, Chappell K, Zou B. Medical errors-what and when: what do patients want to know? Acad Emerg Med. 2002;9:1156–61.
12Association of American Medical Colleges (AAMC). Patient Safety and Graduate Medical Education. Vol 2003. AAMC, 2003.
13Pierluissi E, Fischer MA, Campbell AR, Landefeld CS. Discussion of medical errors in morbidity and mortality conferences. JAMA. 2003;290:2838–42.
14Orlander JD, Barber TW, Fincke BG. The morbidity and mortality conference: the delicate nature of learning from error. Acad Med. 2002;77:1001–6.
15Leapfrog Group. Safety Practices. Vol 2005. Leapfrog Group, 2005.
16National Quality Forum. Mission Statement. Vol 2005. National Quality Forum, 2005.
17National Patient Safety Foundation. Mission/Goals. Vol 2005. National Patient Safety Foundation, 2005.
18Reason J. Human error: models and management. BMJ. 2000;320:768–70.
19Hoff T, Jameson L, Hannan E, Flink E. A review of the literature examining linkages between organizational factors, medical errors, and patient safety. Med Care Res Rev. 2004;61(1):3–37.
20Ferlie EB, Shortell SM. Improving the quality of health care in the United Kingdom and the United States: a framework for change. Milbank Q. 2001;79:281–315.
21Shortell SM, Jones RH, Rademaker AW, et al. Assessing the impact of total quality management and organizational culture on multiple outcomes of care for coronary artery bypass graft surgery patients. Med Care. 2000;38:207–17.
22Weiner BJ, Alexander JA, Shortell SM. Leadership for quality improvement in health care; empirical evidence on hospital boards, managers, and physicians. Med Care Res Rev. 1996;53:397–416.
23Weiner BJ, Shortell SM, Alexander J. Promoting clinical involvement in hospital quality improvement efforts: the effects of top management, board, and physician leadership. Health Serv Res. 1997;32:491–510.
24Croskerry P, Sinclair D. Emergency medicine: a practice prone to error. Can J Emerg Med. 2001;3:271–76.