Increasing awareness of the magnitude of adverse events due to medical error1–5 has engendered a discourse on their reporting and disclosure. It has long been accepted that physicians have an ethical duty to disclose medical errors to patients and families.6 Many professional societies have established guidelines,7–9 which generally recommend that disclosure should consist of at least three components: the truth about the incident, an apology, and an assurance that measures will be put in place to prevent a recurrence.7,8,10 Disclosure in pediatrics has been subject to focused inquiry because of the unique role of the parent or caregiver.11,12 In a study in which parents were asked about thresholds for disclosure, parents wanted to be informed of errors regardless of their severity.13 Parents may also want to be informed of near misses,14 perhaps because they value this information as protectors and advocates for their children.
Despite available guidelines, research has begun to reveal uncertainties and gaps between recommendations and practice.15–20 We know that there are discrepancies between what patients want and what physicians are willing to do; physicians “choose their words carefully” and are reluctant to provide emotional support to patients in spite of acknowledging personal emotional distress in the wake of an error.10,21 There is evidence that physicians and medical trainees feel ill-equipped to perform disclosure22 and to deal with the emotional consequences.23
Although including disclosure of adverse events as part of formal undergraduate or postgraduate medical curricula would provide the rationale for and knowledge of the process and mechanics of disclosure, performance of the skill by trainees might be constrained by social structures within the teaching setting.24–27 The objectives of this study were to explore pediatric residents' knowledge, attitudes, and self-reported behaviors with respect to disclosure. To build an educational program around disclosure, we sought complementary information through both questionnaires and focus groups.
This mixed-methods study used self-administered questionnaires and focus groups.
Population and setting
Participants were first- through fourth-year pediatric residents at the Hospital for Sick Children, a freestanding tertiary children's hospital affiliated with the University of Toronto. Canadian pediatric residency involves three years of general pediatric training followed by either a fourth year of general pediatrics (with responsibilities including the teaching, clinical, and administrative leadership roles of a chief resident) or the first year of a subspecialty fellowship. This study included only general pediatric residents (i.e., rotating residents from pediatric neurology or metabolics/genetics, etc., were not included). Residents on leave at the time of the study were not eligible.
In the fall of 2006, after having sent an e-mail announcement, we delivered a 56-item questionnaire, based on that published by Garbutt et al11 to residents' mailboxes. The questionnaire included questions regarding general attitudes about medical errors, defining errors, communicating about errors with families, deciding whether to disclose, barriers and fears around disclosing, emotional impact, and experience with and training in disclosure. Response formats ranged across open-ended numeric or text responses, four-point Likert scale, multiple-choice, and “choose all that apply.” A $5 coffee card was included with the questionnaire. Questionnaires were completed anonymously and returned to residency administrative personnel, who entered respondents into a draw for a gift certificate. Two e-mail reminders were sent to nonresponders.
Focus groups, conducted during a scheduled weekly educational half-day, explored and documented residents' knowledge, attitudes, and experiences regarding disclosure of adverse events. We mailed invitations to participate to all core pediatric residents. To explore potential differences at different levels of training, we conducted three focus groups: one each for first-year, second-year, and combined third-/fourth-year residents. Because of a scheduling conflict, two fourth-years attended the second-year group. Each participant gave written consent. One of the study investigators (K.T.), who was unknown to the residents, facilitated the groups using a semistructured interview guide. The focus groups were audiotaped.
For descriptive data from the survey, we determined proportions, means, and medians. Our analysis of the focus group data followed grounded theory methodology28,29; all four investigators and a research assistant reviewed verbatim transcriptions of the focus groups to identify major themes. We then reread the transcripts and manually coded for the presence of each theme, selecting quotations to exemplify the themes. Confidentiality was maintained by removing all identifiers from transcript records and study reports. The study was approved by the research ethics board at the Hospital for Sick Children.
Out of 64 eligible residents, 37 (58%) completed questionnaires, including 11 of 18 first-years, 12 of 18 second-years, 8 of 20 third-years, and 6 of 8 fourth-years. Of those 37, 30 (79%) agreed with the statement, “Medical errors are one of the most serious problems in health care.” The residents' estimates of the rates of adverse events in hospitalized pediatric patients were highly varied. For severe adverse events, the median estimate was 3% (range 1%–40%), for minor adverse events, 20% (range 0%–60%), and for near misses, 20% (range 0%–60%). The median estimate of the percentage of Canadian pediatricians likely to be sued for malpractice in the next year was 5% (range 1%–15%); the median estimate of the likelihood of being named personally in a lawsuit was 1% (range 0%–10%). Twenty-six of the 37 (70%) agreed that most errors are caused by system failures rather than individuals.
The major findings of the remaining questions are summarized in Table 1. Almost all respondents agreed that errors should be disclosed to families, but there was less agreement about near misses. A minority feared disclosing an error would damage a family's trust in their competence. Most agreed that disclosing a serious error would be difficult. The reasons chosen most frequently for not disclosing were if the family would not want to know (11/37; 30%) or would not understand (15/37; 41%).
The residents next reviewed the following clinical scenario.
A patient presents with symptoms of a urinary tract infection. You send a urine culture and start antibiotics but forget to follow up on the sensitivities, which show the antibiotic given is not effective. The patient becomes sicker and subsequently requires admission for intravenous antibiotics.
Nearly all identified this as an error, but only a minority would “definitely” disclose it, with many saying they “probably” would. Almost all would be upset and concerned by how it would damage their reputation, but fewer would fear being sued. Most would be willing to give details, to say “sorry,” and to mention how the error would be prevented in the future; fewer than a third were willing to say they made a “mistake” or acknowledge the child was harmed.
When asked about their own experiences, almost all reported personal involvement with at least one error. Most had performed or witnessed a disclosure, and most (28/37; 75%) were satisfied with these disclosures. A minority had received education on how to disclose, the timing of which was evenly split between medical school and residency. All were interested in receiving education/training, as well as coaching, from an expert after an error had occurred.
Most (26/37; 71%) were not confident that health care organizations adequately support physicians in coping with the stress of medical errors and agreed that errors had negatively impacted their confidence and anxiety levels. Almost all were interested in counseling if involved with a serious error.
Almost all of the residents (35/37; 96%) agreed that physicians should report errors to the organization; 36/37 (97%) agreed physicians should discuss errors with colleagues, and most had told a chief resident or staff physician about an error. However, only a minority had delegated or completed a written or online report, and none had told a division chief or department chair. Responses were similar when asked about their future plans for reporting. Thirty-three of 37 (89%) agreed that system changes occur after errors are reported, and 35/37 (96%) agreed that physicians need to know about errors in the organization in order to improve patient safety. Most reported receiving this information in rounds (e.g., M&Ms) and informal discussions, whereas very few cited the hospital's patient safety program.
Focus groups included 6 (of 18 eligible) first-year residents, 12 (of 18) second-year residents, and 6 (of 28) third- and fourth-year residents. Analysis of the qualitative data focused on residents' experiences with medical errors to gain additional insight into how they understood these events, how they had acted in these situations, what their concerns were regarding their competence in dealing with disclosure, and how those concerns might be addressed through education and training. We used thematic coding to identify the main themes residents used to describe their experiences and axial coding by group to analyze the themes both within and between groups of residents. Consistent with grounded theory, we began by identifying specific errors reported by residents. Using theoretical questioning and comparison, we conducted a microanalysis of the way each of the errors was described to identify themes. Theoretical questioning is a technique intended to assist in coding by revealing in more depth the properties and dimensions of specific themes. For example, we identified the nature of the errors as a category and, through theoretical comparisons among errors, noted that the way errors were described differed in terms of how responsibility was allocated, the degree of harm that resulted, and social context. Errors were then selectively coded for responsibility type (self, other physician, other health care worker, or system), for degree of harm (minor or serious), and for the actions and experiences residents reported. In addition, we analyzed and identified themes from discussions regarding factors that either facilitated or inhibited disclosure. Finally, we used axial coding to explore the way errors were described across groups (e.g., did the themes differ between first-, second-, third-, and fourth-year residents, and if so, how?).
The main themes that emerged from the grounded theory analysis were (1) “causal mapping” or assignment of responsibility for errors, (2) specific characteristics of errors, and (3) the social context in which the participant observed or experienced errors. Social context was further analyzed and coded to identify both the ways in which residents situated themselves in relation to errors they reported (e.g., proximate or distant) and how social context affected their decision about what, to whom, and how to disclose. Although we conceived this study to address disclosure, we discovered that residents' thoughts and feelings around disclosure of adverse events to patients and families were closely intertwined with their thoughts and feelings around reporting those events within the institution. Thus, we analyzed how social context affected their actual experiences and shaped their future intentions for both disclosing and reporting.
The data from the focus groups are presented in Table 2. As this table illustrates, participants in Focus Groups 1 and 3 reported a similar number of errors, whereas the number reported by Focus Group 2 was dramatically lower. This finding is noteworthy because Focus Groups 1 and 3 were homogeneous cohorts of peers, whereas Focus Group 2 included a large cohort of second-years and two fourth-years. The presence of the two fourth-year residents appeared to have had a significant inhibiting effect on the willingness of second-years to discuss personal experiences with errors, and this was supported by the group dynamics observed by the investigator facilitating the focus groups. Whereas Focus Groups 1 and 3 engaged in lively discussions of personal experiences, the discussions in Focus Group 2 included only two specific examples of errors and were much more theoretical in nature.
Participants in Focus Group 2 raised significantly fewer specific examples of errors and, based on an analysis of the number of interventions by the facilitator and average length of the intervention, these participants also appeared far less engaged and willing to volunteer information. As Table 3 indicates, the number and average length of interventions by the facilitator in Focus Group 2 was almost triple the number and average length of her interventions in Focus Groups 1 and 3. Although there may be a number of factors affecting the willingness of participants in the focus groups to discuss errors, these data suggest that social context, that is, the size and composition of the group, may also have an impact.
Although participants in Focus Groups 1 and 3 were more willing to discuss specific errors, the analysis presented in Table 2 suggests that these discussions emphasized errors that were “socially distant” or in which responsibility was diffused throughout the system. Twenty-six errors were raised across the groups; however, only four participants raised errors for which they were personally responsible, and none of these had serious consequences. Thus, participants' descriptions of errors were analyzed to examine the extent to which errors were “claimed” or located in the current context and experience of residents versus those errors that were situated as distant to participants. The following quote illustrates the difference between two ways in which errors were described, the first in which the participant situates him/herself as proximate to the situation, versus the second in which the participant distances him/herself from the error:
An interesting situation for me was a patient with a … rheumatology condition with the joints and I was taking care of him … and I ordered one dose of medication … he got one dose of the wrong dose—higher than it should have been, and the child was fine … and I caught it and changed the dose. But I felt it was my responsibility to tell the family and I did. - —Focus Group 3, emphasis added for illustration
The participant clearly feels socially and contextually close to the incident, using the pronoun “I,” describing the situation as local and relatively recent, and claiming responsibility for the error, albeit a minor one with no significant harm.
As a medical student, I had an experience where I had a patient who came into a community hospital in Northern … Ontario…. The staff doctor that was with her was pretty uncomfortable because it wasn't her that had missed it. It was an outside clinic. It was a doctor that she didn't know, so she said this is the worst situation you could be in because it's not my mistake and it's hard for me to apologize for it, and I don't want to get somebody else in trouble because I don't know what they were thinking at the time, but it's obviously a mistake. - —Focus Group 1, emphasis added for illustration
This quote illustrates a report of an error that is socially and contextually distant from the participant who raised it. The incident did not occur at the hospital where the focus groups took place and was at a community rather than academic institution. The error was made by an “outside clinic” and occurred in the past, and the participant, as a student, was an observer rather than a key participant in the error. As Table 2 indicates, although there was a dramatic difference in the numbers of errors raised by Focus Group 2 and the other two groups, in general, no participant in any group claimed responsibility or even described an error with serious consequences in a socially and contextually proximate way. Table 2 suggests that, as residents gain more experience, they do seem more willing to situate errors in more contextually and socially proximate ways, at least in a “safe” environment where confidentiality is ensured and only peers are present.
It is worth noting that residents described certain contextual and social factors as both facilitating and inhibiting disclosure (illustrative quotes provided in Table 4). For example, some saw hierarchy as protecting junior doctors, and therefore facilitating disclosure by some residents, whereas others highlighted how hierarchy inhibits residents from reporting errors committed by attending physicians. As suggested by the different outcomes of the focus groups themselves, residents reported that it was much easier to disclose errors to “insiders” such as peers than to “outsiders” such as families or professionals from other disciplines or institutions. Residents noted how formal reporting, the seriousness of the consequences of the error, and past experiences with reporting of adverse events each influenced disclosure. Level of experience, also a key theme, was interpreted differently by different participants. Whereas some suggested that being “junior” made disclosure less risky, others were concerned that disclosing an error as a junior person was particularly risky to one's reputation. Overall, the data illustrated that “system” errors, for which isolating responsibility is difficult, were easier to report; these were often disclosed as “team errors.” Finally, the quality of relationships among persons involved in the situation was also identified as both a facilitator and inhibitor of reporting. For example, “open” and “trusting” relationships with other professionals or families facilitated disclosure; poor-quality relationships inhibited disclosure.
Discussion and Conclusion
This study both reinforces and extends findings of previous studies regarding disclosure of adverse events. As a mixed-methods study, it combines different, complementary types of information. Quantitative questions (i.e., yes/no questions and numeric ratings) could be asked in a survey, and qualitative information could be gleaned from open-ended questions and direct observation of social dynamics.
Of the participating residents, 92% agreed that adverse events should be disclosed in general. However, when shown a scenario (which 97% correctly identified as an adverse event), only 43% said that they would definitely disclose it. One might assume that this gap is simply due to lack of knowledge or comfort, but our qualitative analysis suggests that residents may also be deciding whether or not to disclose according to circumstances around the adverse event.
Both the quantitative and qualitative results indicate that residents would find it difficult to disclose an error. Concern about losing the patient's or family's trust may inhibit disclosure. Not being instructed how to disclose errors may also contribute to the difficulty. In our study, more residents had performed disclosure than had had the opportunity to observe the skill. The skills required to disclose adverse events are beginning to be investigated, but disclosure could be considered a variation of the well-studied skill of “breaking bad news.” The literature on breaking bad news reveals that although physicians are expected by patients to deliver the news with understanding, empathy, reassurance, and flexibility, they may not have the skills to do so, may not know how to deal with patients' reactions, or may not be prepared to deal with emotions it sparks within themselves.30 The lack of adequate training in breaking bad news is recognized as a problem in medical education, and it may be even worse in the pediatric setting.31 Furthermore, disclosure of adverse events may be more problematic than breaking bad news in general; owning up to an error is certainly more fraught with the health care provider's own emotions.23 Clearly, the idea of being involved in a medical error upset the residents in this study; most agreed that errors had hurt their confidence and raised their anxiety.
A key new theme that emerged from this study was the importance of context—particularly social context—in the experiences residents have and the decisions they make when faced with adverse events. Trainees' positions within the larger social context appear to be the major driver of whether and how they disclose an error. Whether they made an error themselves or discovered someone else's error was paramount; there was tacit agreement that another person's error is his or hers to disclose.
This study has some limitations. It was carried out in a single academic, tertiary care institution. However, the University of Toronto's pediatric residency program is the largest in Canada and has included graduates of all of the Canadian medical schools, suggesting that our results may generally apply to Canadian pediatric residents. Furthermore, because the Royal College of Physicians and Surgeons of Canada has requirements very similar to those of the Accreditation Council for Graduate Medical Education and, therefore, knowledge and attitudes around disclosure in this group of Canadian-trained residents are similar to those of their American counterparts, these results, although exploratory, may be somewhat generalizable to American pediatricians and residents as well.11
Although it would be interesting to analyze the variation in survey responses by postgraduate year, our sample size did not yield sufficient statistical power to do so. Although we tried to create safe environments in the focus groups by dividing participants by year, we achieved this in only two out of the three groups. Ironically, although the presence of both second- and fourth-year residents in Focus Group 2 may have inhibited the discussion of errors, this dramatic effect sensitized us to the importance of context in disclosure. These findings regarding social context require additional study to clarify how much they affect behavior in real settings and across other institutions and contexts.
Although preliminary, these findings suggest that both educators and trainees should recognize the contextual and social factors that influence the willingness to disclose or report an error. It may be valuable to develop training that explicitly addresses the skills required to disclose errors in ways that are socially feasible and constructive. Furthermore, the themes generated in our study could be used to generate a framework for predicting the difficulty of disclosing in certain circumstances and for building institutional supports, such as hospital guidelines, that facilitate, rather than inhibit, the disclosure of adverse events.
These findings add to our understanding of facilitators and inhibitors of AE disclosure and reporting. The influence of social context on this area warrants further study and should be considered in medical curriculum design as well as implementation of hospital guidelines regarding disclosure and reporting.
The authors wish to thank Dr. Tom Gallagher and colleagues for permitting the use of a modified version of their questionnaire prior to its publication.
This study was supported by an educational research grant from Pediatric Consultants at the Hospital for Sick Children.
This study was approved by the research ethics board of the Hospital for Sick Children.
The abstract of an earlier version of this article was presented at the 2009 meeting of the Canadian Pediatric Society and the Ninth Annual Canadian Healthcare Safety Symposium.
1 Brennan TA, Leape LL, Laird NM, et al. Incidence of adverse events and negligence in hospitalized patients. Results of the Harvard Medical Practice Study I. N Engl J Med. 1991;324:370–376.
2 Baker GR, Norton PG, Flintoft V, et al. The Canadian Adverse Events Study: The incidence of adverse events among hospital patients in Canada. CMAJ. 2004;170:1678–1686.
3 Wilson RM, Runciman WB, Gibberd RW, Harrison BT, Newby L, Hamilton JD. The Quality in Australian Health Care Study. Med J Aust. 1995;163:458–471.
4 Institute of Medicine Committee on Quality Health Care in America. To Err Is Human: Building a Safer Health System. Washington, DC: National Academy Press; 1998.
5 Vincent C, Neale G, Woloshynowych M. Adverse events in British hospitals: Preliminary retrospective record review. BMJ. 2001;322:517–519.
6 American Medical Association Council on Ethical and Judicial Affairs. Code of Medical Ethics Annotated Current Opinions. 2004–2005 ed. Chicago, Ill: American Medical Association; 2004.
7 When Things Go Wrong: A Consensus Statement of the Harvard Hospitals. Boston, Mass: Full Disclosure Working Group; 2006.
8 Communicating With Your Patient About Harm: Disclosure of Adverse Events. Ottawa, Ontario, Canada: Canadian Medical Protective Association; 2008.
9 National Quality Forum. Safe Practices for Better Healthcare 2006. Washington, DC: National Quality Forum; 2007.
10 Gallagher TH, Waterman AD, Ebers AG, Fraser VJ, Levinson W. Patients' and physicians' attitudes regarding the disclosure of medical errors. JAMA. 2003;289:1001–1007.
11 Garbutt J, Brownstein DR, Klein EJ, et al. Reporting and disclosing medical errors: Pediatricians' attitudes and behaviors. Arch Pediatr Adolesc Med. 2007;161:179–185.
12 Loren DJ, Klein EJ, Garbutt J, et al. Medical error disclosure among pediatricians: Choosing carefully what we might say to parents. Arch Pediatr Adolesc Med. 2008;162:922–927.
13 Hobgood C, Tamayo-Sarver JH, Elms A, Weiner B. Parental preferences for error disclosure, reporting, and legal action after medical error in the care of their children. Pediatrics. 2005;116:1276–1286.
14 Matlow AG, Moody L, Laxer R, Stevens P, Goia C, Friedman JN. Disclosure of medical error to parents and paediatric patients: Assessment of parents' attitudes and influencing factors. Arch Dis Child. 2010;95:286–290.
15 Mazor KM, Simon SR, Gurwitz JH. Communicating with patients about medical errors: A review of the literature. Arch Intern Med. 2004;164:1690–1697.
16 Gallagher TH, Levinson W. Disclosing harmful medical errors to patients: A time for professional action. Arch Intern Med. 2005;165:1819–1824.
17 Levinson W, Gallagher TH. Disclosing medical errors to patients: A status report in 2007. CMAJ. 2007;177:265–267.
18 Levinson W. Disclosing medical errors: We're halfway there. Arch Pediatr Adolesc Med. 2008;162:991–992.
19 Gallagher TH, Waterman AD, Garbutt JM, et al. US and Canadian physicians' attitudes and experiences regarding disclosing errors to patients. Arch Intern Med. 2006;166:1605–1611.
20 Gallagher TH, Studdert D, Levinson W. Disclosing harmful medical errors to patients. N Engl J Med. 2007;356:2713–2719.
21 Gallagher TH, Garbutt JM, Waterman AD, et al. Choosing your words carefully: How physicians would disclose harmful medical errors to patients. Arch Intern Med. 2006;166:1585–1593.
22 White AA, Gallagher TH, Krauss MJ, et al. The attitudes and experiences of trainees regarding disclosing medical errors to patients. Acad Med. 2008;83:250–256.
23 Waterman AD, Garbutt J, Hazel E, et al. The emotional impact of medical errors on practicing physicians in the United States and Canada. Jt Comm J Qual Patient Saf. 2007;33:467–476.
24 Lee SK, Cowie SE. Medical students and remediation of error. JAMA. 2001;286:1082–1083.
25 Wusthoff CJ. Medical mistakes and disclosure: The role of the medical student. JAMA. 2001;286:1080–1081.
26 Fost N. Ethical issues in whistleblowing. JAMA. 2001;286:1079.
27 Rajendran PR. Ethical issues involved in disclosing medical errors. JAMA. 2001;286:1078.
28 Strauss AC, Corbin J. Basics of Qualitative Research: Techniques and Procedures for Developing Grounded Theory. Thousand Oaks, Calif: Sage; 1998.
29 Kuper A, Reeves S, Levinson W. An introduction to reading and appraising qualitative research. BMJ. 2008;337:a288.
30 Phipps LL, Cuthill JD. Breaking bad news: A clinician's view of the literature. Ann R Coll Physicians Surg Can. 2002;35:287–293.
© 2010 Association of American Medical Colleges
31 Dube CE, LaMonica A, Boyle W, Fuller B, Burkholder GJ. Self-assessment of communication skills preparedness: Adult versus pediatric skills. Ambul Pediatr. 2003;3:137–141.