The CPSO and CMPA need to PERMIT and PROMOTE a Just Culture of Safety.

Talking with Patients about Other Clinicians' Errors

Thomas H. Gallagher, M.D., Michelle M. Mello, J.D., Ph.D., Wendy Levinson, M.D., Matthew K. Wynia, M.D., M.P.H., Ajit K. Sachdeva, M.D., Lois Snyder Sulmasy, J.D., Robert D. Truog, M.D., James Conway, M.A., Kathleen Mazor, Ed.D., Alan Lembitz, M.D., Sigall K. Bell, M.D., Lauge Sokol-Hessner, M.D., Jo Shapiro, M.D., Ann-Louise Puopolo, B.S.N., R.N., and Robert Arnold, M.D.

N Engl J Med 2013; 369:1752-1757October 31, 2013DOI: 10.1056/NEJMsb1303119

You are a young neurologist practicing in a small hospital. You admit a 55-year-old woman with hypertension and type 2 diabetes mellitus who had an embolic stroke at home. On reviewing the patient's medical record, you notice that she appears to have been in atrial fibrillation during two electrocardiographic (ECG) tests during visits to the office of her primary care physician (PCP) for palpitations. Her PCP, an internist who provides many of your referrals, read both ECGs as normal and attributed her palpitations to “probable mitral-valve prolapse and anxiety.” The patient is currently in normal sinus rhythm. You show the internist the ECGs and express concern that they indicate atrial fibrillation. He politely disagrees and says you are confused by noise from his old ECG machine. However, when you ask two cardiologists to look at the ECGs, both immediately say “A-fib.” The internist requests that you transfer the patient to his service (see the video, available with the full text of this article at

Although a consensus has been reached regarding the ethical duty to communicate openly with patients who have been harmed by medical errors,1-6 physicians struggle to fulfill this responsibility.7-10 One particular challenge is that although the literature assumes the physician providing the disclosure also committed the error, health care today is delivered by complex groups of clinicians across multiple care settings.11 In addition, safety experts emphasize the role that system breakdowns play in adverse events.12 Thus, many decisions about discussing errors with patients involve situations in which other clinicians were primarily responsible for the error.13

Confronting the apparent error of a colleague raises challenging questions about whether an error occurred, how the error arose, which professionals carry what responsibilities, and how to talk with the patient about the event.4,14,15 Existing guidelines emphasize the overall importance of disclosing errors, but (with the exception of the case study of the American College of Physicians Ethics and Human Rights Committee)16 they offer little guidance on disclosing others' mistakes; this lack of guidance heightens clinicians' uncertainty about what to do. Consequently, patients may be told little about these events, and opportunities to build trust, ensure that learning occurs after errors, and avoid litigation may be lost.17,18

We convened a working group of experts in patient safety, medical malpractice insurance and litigation, error disclosure, patient–provider communication, professionalism, bioethics, and health policy. After the meeting, a subgroup of attendees collaborated to refine these concepts and draft this manuscript. Below, we describe recommendations that extend existing guidelines for clinicians and institutions on communicating with patients about colleagues' harmful errors.

Challenges When It Is Not “My Error”

The rationales for disclosing harmful errors to patients are compelling and well described.19,20 Nonetheless, multiple barriers, including embarrassment, lack of confidence in one's disclosure skills, and mixed messages from institutions and malpractice insurers, make talking with patients about errors challenging.21 Several distinctive aspects of disclosing harmful errors involving colleagues intensify the difficulties.

One challenge is determining what happened when a clinician was not directly involved in the event in question. He or she may have little firsthand knowledge about the event, and relevant information in the medical record may be lacking. Beyond this, potential errors exist on a broad spectrum ranging from clinical decisions that are “not what I would have done” but are within the standard of care to blatant errors that might even suggest a problem of professional competence or proficiency.22

One potential solution to this lack of information is to talk with the involved colleague or colleagues about what happened, whether it was a harmful error, and what, if anything, to tell the patient. In practice, however, fear of how a colleague will react, along with strong cultural norms around loyalty, solidarity, and “tattling”23 may deter such conversations. There is a natural reluctance to risk acquiring an unfavorable reputation with colleagues, disrupting relationships among and within care teams, or harming one's institution. Power differentials, including those associated with seniority, sex, and race, previous relationships with colleagues, interprofessional and other cultural differences, and in some cases, dependence on colleagues for referrals all create complicated interpersonal dynamics.24 Pragmatically, time constraints and coordinating meetings with multiple clinicians pose additional barriers.

Clinicians might be tempted to use the patient's medical record to raise concerns about a potential error without initiating a direct conversation. Although this approach can avoid awkwardness and maintain the appearance of collegiality, it arguably transgresses the norm of loyalty even more than a direct conversation, since it can create evidence for a malpractice suit without allowing the colleague to dispel misconceptions.

Although health care institutions could help determine what happened and plan for disclosure, some clinicians will consider turning to their institution to be problematic. They may worry that reporting a concern to the institution might lead to an unpredictable, punitive cascade — or, on the other end of the spectrum, that no action will be taken.25 The clinicians and institutions involved may have different malpractice insurers that disagree about how to handle the event. Finally, many clinicians work in small practices without access to institutional resources to help them figure out what happened and navigate the disclosure conversation.

Even when the facts surrounding harmful errors seem clear, other challenges can make it difficult to know what to say to the patient. Clinicians may have legitimate concerns about destroying patients' trust in the involved colleague, especially if there is an ongoing care relationship. There are also worries about triggering litigation. Although some physicians might be willing to subject a colleague to difficult conversations with an angry patient or family, few will find it easy to expose him or her to a potential malpractice suit. Most states protect some aspects of disclosure conversations from use in litigation, but this protection is incomplete and might not extend to protecting an unrelated third party to the disclosure.26 And although research suggests that good communication about adverse events may reduce lawsuits,27,28 data are lacking from studies to indicate how to disclose others' errors while minimizing the risk that a patient will initiate a claim.

Where Do We Go from Here?

The approach to communicating with patients about other clinicians' errors should be determined through research into how this challenge arises; the preferences of patients, clinicians, and institutions regarding handling such situations; and outcomes data regarding disclosure strategies. The following principles should be refined as data and experience accumulate.

Patients and Families Come First

Although anxieties about damaging collegial relationships loom large in situations of potential error involving other clinicians, a patient's right to honest information shared with compassion about what happened to him or her is paramount. Simply put, when disclosure is ethically required, the fact that it is difficult must not stand in the way. Patients and families should not bear the burden of digging for information about problems in their care.

It must also be acknowledged that many families will need financial help after a serious error and will have a hard time accessing compensation without information about what happened. Clinicians rightly perceive the current medical liability system as flawed and understandably worry that they may not be treated fairly should a patient file a claim.29 But these concerns do not obviate clinicians' duty to be truthful with patients; as professionals, clinicians are expected to put the patient's needs above their own.

Explore, Do Not Ignore

Before initiating a disclosure conversation about a colleague's possible error, a clinician's first obligation is to obtain the facts. Patients' interests are not served by communicating inaccurate or speculative information, and colleagues deserve the chance to correct mistaken assumptions and join disclosure conversations with their patients.

A strengthened commitment by clinicians to “explore, don't ignore” potential errors is needed, and it will require that clinicians improve their ability to discuss quality issues with one another. This commitment is fundamental to the self-regulation that lies at the heart of medical professionalism. Professional self-regulation should not be conceived of as something individual clinicians do, but rather as something the profession does collectively — and can only do by sharing and acting on information together.30

Ideally, we envision the process starting with a colleague-to-colleague conversation about what happened. Interacting directly with the involved colleague is part of our professional responsibility. It is how we would hope a colleague would treat us, and it can promote learning. For a productive discussion to occur, it is essential to frame the conversation in ways that minimize a colleague's defensiveness. A shift to a more proactive approach to discussing colleagues' potential errors should be balanced by a willingness not to rush to judgment. Explorations should be undertaken with the assumption that persons who were not directly involved in the care have incomplete information, and the discussions should be approached with curiosity rather than accusations.31

The goal of the discussion with the involved colleague is to establish what happened and, if needed, how to communicate with the patient. The path forward will depend on the outcome of the peer-to-peer conversation. The colleagues may agree there was no harmful error, and the process can stop. If they agree there was a harmful error, they can discuss what needs to be reported through institutional channels and disclosed to the patient. The colleagues may also disagree about what happened or whether disclosure is warranted. When these or other challenges described below occur, it is appropriate to turn to the institution or health care organization for assistance, if possible.

The challenge of disclosing another providers' error can arise in various situations. Table 1Table 1Disclosing Harmful Errors in Common Situations Involving Other Clinicians. outlines several common situations and proposes a disclosure strategy for each. The recommendations place a priority on patients receiving needed information about harmful errors through skillfully executed disclosure conversations. The recommendations regarding who is responsible for the disclosure were derived by considering who has the strongest ongoing relationship with the patient, the best understanding of what happened and its implication for the patient, responsibility for the patient's current care, and the most experience with disclosure in complex situations (such as those involving multiple institutions). The proposed strategies also recognize the advantages of the disclosure being conducted jointly by the involved clinicians. This avoids sending the patient mixed messages, ensures that key information is communicated clearly (rather than merely hinting at the error, so that the patient is left responsible for “connecting the dots”), and demonstrates shared responsibility for transparency.

Institutions Should Lead

Although colleague-to-colleague discussions should be the starting point for exploring potential errors, institutions are ultimately responsible for ensuring that high-quality disclosure conversations occur with patients, regardless of which clinicians were involved in the event.4 Institutional leadership is especially important when the patient had considerable harm, multiple clinicians or other institutions were involved, communication among colleagues has broken down, the colleagues disagree about what happened or whether disclosure is warranted, and concerns are raised about conflict of interest (e.g., the colleague in question is a financial competitor). Institutions that play a prominent role in such situations ensure that a careful review of the event is performed and that clinicians have not assumed that disclosure is someone else's responsibility and left the patient in the dark.

Institutions should support conversations between clinicians as they seek to explore potential errors. Many institutions are developing just-in-time disclosure coaching programs that could help clinicians conduct respectful conversations with colleagues about potential errors.15 A disclosure coach can facilitate peer-to-peer discussions, enabling each party to share his or her perspectives with less defensiveness. Role modeling by senior colleagues is also essential to encourage meaningful discussions among clinicians.

Existing formal venues for discussing concerns about quality such as morbidity-and-mortality conferences and peer-review committees could also address questions about potential errors involving colleagues. In addition, less formal mechanisms such as a “curbside consult” with a quality expert or risk manager could help clinicians access the institution's event analysis expertise (under the appropriate peer-review and quality-improvement privileges) while minimizing clinicians' fear of a punitive review process. Institutions could enhance their preparation for handling errors involving colleagues by using the atrial fibrillation case to simulate how existing quality, safety, and risk structures would help clinicians respond. Institutions could also use the case to identify opportunities for improvement. Such organizational preparation is preferable to confronting these crises only when they arise.

Institutions should also strengthen “just cultures,” which are “atmospheres of trust in which people are encouraged, even rewarded, for providing essential safety-related information — but in which they are also clear about where the line must be drawn between acceptable and unacceptable behavior.”32-34 Just cultures encourage clinicians to report adverse events and help address hierarchy issues involving nurses and trainees that can obstruct the free flow of information to the patient.

Similar resources will be needed for clinicians who do not have a strong institutional connection. These clinicians' liability insurer or insurers could provide similar support and, indeed, many insurers have expanded their disclosure coaching resources. Support could also be provided by local medical societies and national professional organizations. Patient-safety organizations could fill this role over time, have the advantage of strong statutory protections for the confidentiality of information reported to them, and can help bridge the gap in cases that involve multiple institutions.35

What Should the Neurologist Do?

The neurologist in our case is in an awkward position. She is confident that the patient's internist did not diagnose atrial fibrillation, that this error probably contributed to the patient's stroke, and that disclosure to the patient is vital. The internist has rebuffed her without assuaging her concerns. The neurologist's next step should be to tell the internist she plans to request a formal cardiology consultation. With the diagnosis firmly in hand, she should communicate the findings to the internist and attempt to formulate a joint disclosure strategy. If the internist declines or objects to the cardiology consult, the neurologist should seek assistance from the institution's medical director or other senior administrative leader. The neurologist would be well served by support from a disclosure coach.


When faced with a potential error involving another health care worker, our conceptions of professionalism should lead us to turn toward, rather than away from, involved colleagues. Although making the effort to understand what happened and ensure appropriate communication with the patient may challenge traditional norms of collegial behavior and involve additional demands on clinicians' time, transparent disclosure of errors is a shared professional responsibility. Only a collective approach to accountability can fully meet the needs of patients and families after harmful medical errors.

The views expressed are solely those of the authors and do not reflect the official positions of the institutions or organizations with which they are affiliated or the views of the project sponsors.

Supported by a grant from the Greenwall Foundation and by the Risk Management Foundation of the Harvard Medical Institutions.

Disclosure forms provided by the authors are available with the full text of this article at

We thank Robert Hanscom, J.D., Patrice Blair, M.P.H., Beth Cushing, J.D., Brent Tingle, J.D., and Carol Dembe, M.D., J.D., for their important contributions to our working group, Ben Dunlap for project assistance, and the Oregon Medical Association for sharing the video associated with this article.

Source Information

From the Department of Medicine and the Department of Bioethics and Humanities, University of Washington, Seattle (T.H.G.); the Department of Health Policy and Management, Harvard School of Public Health (M.M.M.), the Department of Social Medicine, Harvard Medical School (R.D.T.), the Department of Medicine, Beth Israel Deaconess Medical Center (S.K.B., L.S.-H.), and the Division of Otolaryngology, Brigham and Women's Hospital (J.S.) — all in Boston; the Department of Medicine, University of Toronto, Toronto (W.L.); the Institute for Ethics, American Medical Association (M.K.W.), and the Division of Education, American College of Surgeons (A.K.S.) —both in Chicago; the Center for Ethics and Professionalism, American College of Physicians, Philadelphia (L.S.S.); the Institute for Healthcare Improvement, Cambridge, MA (J.C.); the Department of Medicine, University of Massachusetts Medical School, Worcester (K.M.); COPIC Insurance, Denver (A.L.); CVS Caremark, Woonsocket, RI (A.-L.P.); and the Institute for Doctor–Patient Communication and Section of Palliative Care and Medical Ethics, University of Pittsburgh School of Medicine, Pittsburgh (R.A.).