Communication Failures: An Insidious Contributor to Medical Mishaps

Abstract

Purpose 

To describe how communication failures contribute to many medical mishaps.

Method 

In late 1999, a sample of 26 residents stratified by medical specialty, year of residency, and gender was randomly selected from a population of 85 residents at a 600-bed U.S. teaching hospital. The study design involved semistructured face-to-face interviews with the residents about their routine work environments and activities, the medical mishaps in which they recently had been involved, and a description of both the individual and organizational contributory factors. The themes reported here emerged from inductive analyses of the data.

Results 

Residents reported a total of 70 mishap incidents. Aspects of “communication” and “patient management” were the two most commonly cited contributing factors. Residents described themselves as embedded in a complex network of relationships, playing a pivotal role in patient management vis-à-vis other medical staff and health care providers from within the hospital and from the community. Recurring patterns of communication difficulties occur within these relationships and appear to be associated with the occurrence of medical mishaps.

Conclusion 

The occurrence of everyday medical mishaps in this study is associated with faulty communication; but, poor communication is not simply the result of poor transmission or exchange of information. Communication failures are far more complex and relate to hierarchical differences, concerns with upward influence, conflicting roles and role ambiguity, and interpersonal power and conflict. A clearer understanding of these dynamics highlights possibilities for appropriate interventions in medical education and in health care organizations aimed at improving patient safety.

Medical mishaps are a pervasive problem in health care organizations. In 2000, the Institute of Medicine (IOM)¹ reported that between 44,000 and 98,000 people die every year in U.S. hospitals because of medical errors. The studies on which the IOM relied for its estimates,² as well as other studies of medical error,^3,4 were based on physicians’ reviews of medical records and their judgments of adverse events. Studies in this vein tend to focus on the incidence and nature of errors in medicine and consider errors a function of deficiencies in expertise and training. What is missing from current research is a rich description of the way clinicians in complex organizations such as hospitals experience errors in the context of daily clinical practice.

To be sure, individuals make errors. Yet, studies of organizational accidents in other disciplines suggest that most of the accidents that occur in complex organizations do not usually have a singular cause, but result from a string of latent flaws throughout the system.^5,6 The growing interests in the broader context of medical mishaps and the systems in which health care providers are embedded comes at a time when the conceptual and methodological tools in current use drastically underestimate the role of social, relational, and organizational factors in the generation of adverse medical events.^5–8 With these issues in mind, we undertook this qualitative study to examine the individual resident's experience and perception of the causes and contexts of medical mishaps. Using data collected through a series of 26 semistructured interviews with residents regarding their routine work environments and the medical mishaps with which they had been involved, we explored how communication plays an integral role in many untoward events.

Faulty communication has been implicated both in exhaustively studied and well-known catastrophes such as the explosion of the space shuttle Challenger,^6,9 the release of methocyanate in Bhopal, India,⁶ and in the occurrence of adverse events in other high-risk contexts.^10,11 Communication failures are increasingly being implicated as important latent factors influencing patient safety in hospitals as well.^10,12 A retrospective Australian survey of hospital admissions showed that communication problems were the most common cause of preventable disability or death.¹³ More recently, Chassin and Becher¹² analyzed how it was possible for the wrong patient to undergo an invasive procedure and concluded that “frighteningly poor communication” was a key causal element.

Today, medical care involves shorter hospital stays with a rapid turnover of acutely ill patients. Patient management involves complex investigation and coordination of care by a myriad of medical specialists. Clinical medicine thus involves multiple handoffs with many places where critical information must be effectively communicated. In addition to the horizontal differentiation of labor, vertical divisions of hierarchy and power operate as well.¹⁴ Complex systems are made up of individuals at different hierarchical levels who must constantly interrelate. These relationships are shaped by the relative status of those involved. These structures can have a powerful influence on whether and how critical information is effectively communicated. With this theoretical framework in mind, in this study we sought to understand how social, relational, and organizational structures contribute to communication failures that have been implicated as a large contributor to adverse clinical events and outcomes.

Method

Study Location and Sample

The research site was a 600-bed U.S. teaching hospital with a large graduate medical education program. Our focus was on the inpatient setting. The study was approved by our university human subjects’ committee, and the hospital's Internal Review Board, Quality committee, and department of risk management.

The sample, stratified by specialty, year of residency, and gender, was randomly drawn from a total population of 85 residents within the three specialty residency programs sponsored by the hospital: Surgery, Medicine, and Obstetrics/Gynecology (Ob/Gyn). Although the research site was chosen for convenience, we chose to randomly sample residents to more accurately represent the total population.

The final sample consisted of 26 residents (30% of the population) including five surgery, 17 medicine (11 medicine and six preliminary/transitional), and four Ob/Gyn residents. Fourteen were men and 12 were women. Eleven residents were in their first year of postgraduate training, five in their second year, seven in their third year, two in their fourth year, and one in their fifth year. The disproportionate number of first-year residents in the sample and population was the result of students going on to residencies in other programs after completing the first year. Residents ranged in age from 25–39 years and averaged 29.8 years.

Data Collection and Analysis

We contacted individuals first by letter and then by telephone. We collected data using a semistructured interview approach that allowed residents to respond to themes generated from the literature and incorporated into the interview questions, and also enabled them to reflect on their own unique experiences. Before the interview, respondents were assured of confidentiality and anonymity. Each person signed and returned a confidentiality and voluntary participation agreement and gave us permission to audiotape the interview.

The interview protocol consisted of a series of open-ended questions that focused on the general work environment and the medical mishaps in which each respondent had been involved within the preceding three months. Many terms have been used to describe medical errors but we chose to use the term mishap, a neutral, nonjudgmental term that encompasses a wide range of mistakes from the near-miss to a serious iatrogenic injury, to elicit a broad range of incidents. We asked respondents to describe each mishap in depth and subsequently asked them to categorize each mishap into one or more of six categories (e.g., omission, commission, diagnosis, treatment, medication, or patient management), which we derived from review of existing literature on medical mistakes. We then asked each respondent to describe and categorize the contributory factors associated with each mishap (i.e., elements they thought were linked with the mishap) and to choose the most important factor. We asked additional questions about residents’ work environments and daily activities, how mishaps were acknowledged, and ended with a question about how mishaps could be avoided. Interviews lasted between one to two hours. All interviews were tape-recorded, transcribed, and assigned a unique record number.

Data were analyzed using case analysis techniques suggested by Yin¹⁵ that included putting information into different arrays, creating a matrix of categories and placing evidence within these categories, tabulating the frequency of different events, and conducting an iterative process to build an explanation. Transcribed interviews were entered into a qualitative data computer program to facilitate coding and sorting of interview data and to compile simple counts and statistics once data had been coded. Initial codes were derived from the interview questions and the literature search (see previous paragraph for an example of the six categories for classifying each mishap), or were induced from the narratives. For example, after reviewing all responses to the question “how was the mishap acknowledged,” we found that the answers could be sorted into one of five categories: discussion, written documentation, review at conference, ignored, or unknown. Two researchers coded and sorted interview data into the categories. The two coders agreed in 80% of the cases and all disagreements were resolved through discussion. In addition to the categorical data coding, two researchers read through the interview transcripts repeatedly and then discussed them to induce the set of emerging themes and patterns and build an explanation.

Results

Overview

Respondents reported a total of 70 mishaps, which varied from relatively minor incidents such as a near miss in which two patients’ orders were switched, but subsequently corrected by a nurse, to relatively major incidents, such as a chest tube inserted on the wrong side. Outcomes ranged from no untoward consequences to death. Respondents categorized mishaps most frequently as errors of patient management (29), errors of omission (26), errors of diagnosis (24), and treatment (24), and commission (21). Surprisingly, there were very few medication-related errors (6), probably because this institution has had a computerized order-entry system in place for several years. (The total of all categories exceeds 70 because residents often categorized an incident into more than one category. For example, a case where a patient with cardiomyopathy was sent home from the emergency department with a pneumonia diagnosis was categorized as omission, diagnosis, and patient management.)

Practitioner's knowledge (30) and communication (28) were cited most frequently as the most important factors that contributed to each mishap discussed. On the surface, practitioner's knowledge seems to refer to the extent to which a practitioner's medical knowledge is complete and accurate; yet further content analyses showed that this category was more often used to refer to a practitioner's awareness of certain pertinent information. For example, residents chose this category for information regarding a patient's test results, previous diagnoses/treatments, or other medical historical issues that were not communicated effectively to the practitioner. Thus, we counted instances of lack of information as instances of faulty communication. We found other contributory factors that involved issues of communication/information transmission as well, including factors related to the specific situation (e.g., busy emergency department), work environment (hostile superior), and practitioners’ interpersonal skills. In total, communication failures of one kind or another were an associated or contributory factor in 64 mishaps (91%).

The narratives themselves reveal additional insight. In reflecting on their roles within the hospital, residents described themselves as embedded in a complex network of relationships. The resident plays a pivotal role in patient care and management, vis-à-vis other medical staff and administrators within the hospital as well as community medical practitioners. Although our qualitative analysis suggests that poor transmission and exchange of information accounts for some mishaps, this is only part of the story. Other aspects of faulty communication repeatedly show up within the context of four key relationships and provide a framework for understanding how issues of miscommunication can contribute to many medical mishaps. Representative quotes from the interviews are presented in Table 1. These and other quotes are discussed more fully below.

Table 1: Narratives from Interviews with 26 Residents about Medical Mishaps at a U.S. Teaching Hospital, Classified by Dyads and Communication Themes, 1999–2000

Relational Dyads and Common Communication Failures

Residents and attendings

Faulty communication typically arises in the context of the relationship between a resident and the attending physician with whom the resident works.¹⁶ Several themes are prominent. The attending is both a supervisor and a teacher, whereas the resident performs the bulk of patient care and decision making. We found that residents in our study were concerned about appearing incompetent in front of those with more power and they were hesitant to communicate information that was unfavorable or negative to themselves. A resident does not want to appear ignorant about a patient. He or she wants to appear knowledgeable about the medical condition and about the patient in particular, and to present pertinent but not unnecessary information.

In our study, the hesitancy to communicate information to superiors was evident in situations where a resident did not want to appear incompetent and also in situations where residents thought they may offend those in power. Residents repeatedly commented on their hesitancy to call an attending in the middle of the night. Although it is considered appropriate under certain conditions, unnecessarily waking or bothering an attending is bound to cause some anxiety and even some friction. There is a tension between wanting to be sure one is taking the correct action and wanting to know enough not to have to contact the attending.

Residents’ concerns about offending those in power combined with their perceptions that powerful others would not listen to them or hear their point of view also discouraged residents in our study from productively disagreeing when they had a different point of view. In one case, the attending physician treated an elderly woman with a superficial vein thrombosis with anticoagulants. The resident vehemently disagreed with this course of action, and described why she did not openly object to the treatment:

If I felt like I could actually communicate with that group of attendings I would have tried, but I didn't feel like it would be useful for me. And all it would have done would be to inflame the relations between me and that attending and the patient still would have ended up getting [inappropriate treatment].

The patient had a massive retroperitoneal bleed, was transferred to the intensive care, and survived with serious complications. Although it is not clear which course of action was best in this case, closed lines of communication between the attending and the resident may have impeded optimal patient care.

Another kind of miscommunication between attending physicians and residents concerns the amount of information communicated by superiors. Specifically, residents perceived that attendings provided far too little information to the resident who would be caring for a particular patient.

As one resident said:

The nurses will not take a patient unless they've gotten a report. And they are very strict about that. And you know, doctors simply don't do that. They transfer patients, they do this, they do that, and then they just don't tell you. They don't tell the resident. And, I mean, it's a simple thing but, you know, a five-minute discussion about a patient that's being admitted or being transferred or whatever is worth its weight in gold when something dreadful happens and I'm running down there and I don't even know how old the patient is or what problems they have. They really need to communicate better.

This type of situation occurred between residents and attending physicians within the hospital, and also between residents and consulting physicians from within the hospital and from the larger community.

Hospital residents and community physicians

Lack of information and/or the ineffective communication of information arising from the relationship between hospital residents and community physicians were associated with reported mishaps. This faulty communication occurred most often when the hospital resident admitted a patient who was previously under a community physician's care. At the hospital in which we gathered our data, private physicians may admit a patient whose care is then taken over by the “academic team” of staff physicians and residents. During this complete transfer of patient care, information about the patient is often seriously or completely lacking. As one resident described:

God, we get patients, we're lucky if we get any information when the patients come into the hospital. About half the time we don't even know they're coming. They'll just show up on the floor. There's a huge, huge potential for mishaps to occur that way … between the admitting physician and the people who are ultimately responsible for taking care of the patient in the hospital right then.

Communication failures within this relationship also arise out of role conflict and ambiguity. Occasionally, a private community physician admits a patient to the academic team but then maintains some role in the patient's care. In this context, disagreement about patient care and management, as well as miscommunication, are common. Friction may occur when the resident and the community physician disagree on patient care decisions, or miscommunicate about patient management.^17,18 In one case related by a resident, a 30-year-old woman was admitted by her private physician on a Friday morning and was scheduled for magnetic resonance imaging (MRI) that afternoon. Her physician suspected a dissection in the carotid artery. The resident was aware of the MRI, but had no other information about the patient and did not suspect any particular diagnosis. The resident sent the patient home for the weekend pending the MRI results and was severely chastised by the private physician for doing so against his wishes. The MRI results, which were discovered Monday morning, confirmed dissection in the carotid artery—a potentially serious condition. The resident highlighted some areas of miscommunication:

Our suspicion of the pathology wasn't there because [the patient looked healthy and we didn't know what her physician had in mind]. ... So we asked him [her physician] if we could just get the MRI and then discharge her for him to follow-up on the results. And again, this is me communicating through my intern and then my intern talking to him. And it was fine, we got the MRI and let her go. The communication between our team and, you know, the specialist was such that there must have been a break down in understanding. And I'm wondering if the intern didn't misunderstand the whole thing and maybe I didn't explain it enough to her, and again it's that whole A to B to C to D. You've got to get the information through all those points and there's a break down. I don't know where it happened necessarily. ... It's stressful on the residents and the interns because we're caught in the middle. We can't make our own decisions, and yet the specialists come along whenever they want. So we're either bugging them, you know, calling them, or there's a break down in communication because it's going from the specialists usually to the intern to the resident and the attending. It's like [the game] telephone, you know?

This case illustrates two communication issues. First, little information was communicated to the resident who was asked to carry out the plan of care (the MRI). Second, the communication links were so convoluted that any messages that were conveyed were misinterpreted as they moved through all those involved. In sum, insufficient information, faulty exchanges of existing information, or ambiguous and unclear information seem to characterize incidents involving hospital residents and community physicians. The poor communication often arises out of role conflict and ambiguity—where the boundaries between who has the authority for care and who is responsible for care are unclear. Our evidence suggests that these situations provide fertile ground for mishaps.

Internal medicine residents and specialists

Faulty communication is pervasive in relationships between medicine residents and the residents and/or attendings in different specialties within the hospital. Our analysis revealed some reasons for this. Not surprisingly, the timely and effective exchange of pertinent information was a prominent theme, particularly between the emergency department and internal medicine, as one medicine resident noted:

Once they called me from the ER on a patient who came with congestive heart failure. They didn't mention that the patient had EKG changes and abnormal heart enzymes. She was having a heart attack but we were really super super busy so we didn't have a chance to take a look at the patient until she hit the floor and she was having a heart attack. She was infarcting. And so that's communication. Communication's the problem. When somebody's taking care of the patient, when the patient's transferred from a service to a different service, there should be very good communication for that so that all the information is conveyed there.

In another case, a resident reported her experience with a patient who had just been admitted and transferred to general medicine from the emergency department with dangerously low blood pressure. No one on the floor had been notified of the patient's condition before arrival. The resident to whom the patient was assigned just happened to walk by the room when the patient started to decompensate:

As I walked in, the patient was in relative distress, short of breath, looked sick. And so I asked them to take a blood pressure immediately and the blood pressure was in the seventies over doppler. So I immediately put the patient in Trendelenburg and got some IV fluids going and the patient looked septic to me. And I had to send the patient to the unit immediately. I think that patient should have never come to the [medicine] floor to begin with. And when I look back, the nursing records show[ed] that the patient had been hypotensive in the ER prior to even coming up to the floor.

Fortunately, the outcome in this case was good. The patient was moved to the intensive care unit soon enough and avoided any serious complications, as the resident noted:

We were lucky that we got to the patient soon enough because had the patient been sitting there, as it happens at times, depending on how busy the intern and resident are, and not [been] assessed by anyone, I think we could have had a bad outcome.

We noted earlier how the openness and quality of communication suffers in the presence of private concerns about hierarchy and power. Communication is distorted because of concerns about offending the more powerful party, and also because residents want to avoid conflict or believe that their concern will not be addressed. These issues surface frequently in situations where residents interact with the attendings in other departments, which is often required of them. For example, medicine residents frequently admit patients sent to them by emergency department attendings. When problems arise, a resident may find it difficult and frustrating to speak up, particularly to disagree, with an unknown attending. Residents feel uncomfortable in voicing their objections and perceive that even if they do, they may not get a positive response. One resident in our study described what happened when a resident disagreed with the patient management prescribed by the emergency department:

It's kind of hard for them to accept [disagreement] because they are attendings and we are residents—a lower level of the hierarchy. So it's not that nice—they don't feel that good if they get a call from the resident and they are an attending and that “blah, blah, blah happened and you guys mismanaged that patient.”

There is evidence that role conflict and contrasting core values between different kinds of physicians can influence effective communication.^10,17 These tendencies are exacerbated by the fact that very little direct (face-to-face) communication occurs between primary care physicians and specialist consultants.¹⁸ In our analysis, these kinds of struggles often led to patient management problems. In one case, a cardiologist who was called to consult on a patient put the patient on a medication. The medicine attending took her off the medication. The notes in the patient's record were unclear and there was no verbal communication between the parties. This led to a disagreement with regard to patient management and the resident felt stuck in the middle:

… I guess the attending that was on our team had a different philosophy than the cardiologist who would have kept her on anticoagulants. And, … you know, in his note [he] didn't write it down and never verbalized it to us. I mean, just bad communication.

As this resident went on to explain, much of the communication between consultants and medicine attendings is through written notes in the patient's record—a sometimes ineffective means of communicating:

The charts are the primary means of communication between a lot of different people—the consultants, the interns and everyone. It's not necessarily the best way because a lot of times you can't read the handwriting very well. A lot of times something's forgotten, you're just writing a quick note and they forget to mention something.

Residents and nurses

The relationship between residents and nurses is critical in preventing medical mishaps. Residents make most of the day-to-day decisions about patient care and nurses not only carry out many of the orders, but also are in closest contact with the patient and are better informed about their moment-by-moment condition. Information given to nurses by residents in the form of orders, and information given to residents by nurses in the form of the patient's condition are crucial. Our analysis revealed why both kinds of information may not be communicated effectively.

Residents routinely communicate orders in hospitals by writing or typing the orders in the patient's medical record. Yet, written communication, an impersonal medium with limited capacity for timely feedback,¹⁹ is seldom the most effective way to communicate a plan of action, especially when action needs to be taken quickly. As one resident noted:

I think the main important thing is communication with the nurses. Putting orders in the computer or just putting orders in the chart can be missed and they have been missed before just because people don't look at it. And if you don't tell the nurse on the floor that “this is going to be ordered, watch for it,” like that, it may take them a couple hours to go and get the order from the clerk and do the appropriate work. And I've seen it more than not that that happens—that orders get missed. And just recently there was a chemo treatment order that was missed. It was in the chart early in the morning and it was missed until three o'clock in the afternoon. And I was asking, “Where's the treatment?” and they looked in the chart and said, “Oh nothing was started” because the attending never flagged the chart or didn't tell any of the nurses, so no one looked at it. And then when I was called by the attending in the afternoon [to check] that the chemo started, I asked the nurses and they said no. But there was no nurse to run the chemo at night. So it got delayed until the next morning. That could have been avoided by communication.

Here the delay in treatment occurred without adverse effects, but this type of miscommunication could result in more serious consequences.

Another resident gave a similar opinion and also noted the importance of face-to-face interactions:

I think that communication, especially between the doctors and the nurses who are giving the stuff that you're ordering, is very key. And I've had a lot of positive feedback from the nursing staff that “We appreciate you coming down and telling us this so that it doesn't sit around for three hours and the patient doesn't get their nitroglycerine that they need.” Or whatever.

The personal characteristics of communicators also can influence the character of a communication relationship, sometimes giving rise to interpersonal conflict.^19,20 For example, in one case the resident made a mistake in writing a blood pressure medication prescription for his patient. The nurse noticed this and mentioned it right away, but not in such a way that the resident understood and/or responded. The nurse mentioned it again only after the incorrect dosage was administered:

The nurse had questions and did bring it to my attention, but after the medicine had been given. Now she said that she had asked me as well before she gave it. But she certainly asked me in a much different way after she gave it. So she could have come to me and been a little bit more forceful about how it didn't make sense.

The nonverbal information displayed by one or the other party (communicated through eye gaze, posture, facial expression, and voice tone) can also influence communication,²¹ and our data reflect this. For example, several residents suggested that whether the nurse was “nice” or “aggressive” could make a difference in how an intern responded to a nurse's concerns regarding a patient.

Although personal characteristics may matter, we repeatedly found that communication behaviors are influenced by hierarchy and social structure and, as described below, perceptions that a superior is receptive to receiving information. For example, as one resident related, a 70-year-old-patient who had recently had cardiac bypass surgery was not urinating and had low blood pressure during the night following surgery:

There was a communication problem between the nurse who was taking care of her and myself in that she felt that I was not adequately addressing her concerns about the patient, so she didn't call me back to say that the patient was continuing not to do well … I don't understand. She said she felt intimidated by me or something. It really didn't make a lot of sense to me at the time. ... We actually sat down and had a discussion about this after—with the attending physician and the nurse and myself—about how this had transpired and what had happened. And so that's how I had some insight into her perspective on it. And basically she felt that … I was blowing her off or didn't seem to be approachable and therefore for some reason she felt it was appropriate just not to do anything until the next day.

Although the surgery resident was not certain of the eventual outcome because the patient was transferred to another unit, he reported that there were “potential long-term complications” for the patient due to this mishap.

In the following excerpt from an interview, a resident explained the difficult position nurses often find themselves in.

Resident:

In some cases nurses can be reluctant to question physicians because, you know, because of physicians’ responses to them, questioning them, or residents for that matter.

Primary investigator:

You mean they have a sense that they [physicians] don't welcome criticism, any questioning?

Resident:

Right. So I think this demonstrates that it's very important that nurses feel able to question the doctors...they walk a fine line because they don't always know what's appropriate to question and what's not appropriate to question. ...

Discussion

Our qualitative study provides insight into the insidiousness of faulty communication as a contributor to medical mishaps. The residents in our study perceived that communication difficulties played a role in the vast majority of the medical mishaps they experienced. Our findings are consistent with other research showing a strong link between poor communication, and errors and adverse events.^12,13 Although few studies in medicine have systematically examined this association, the implications drawn from a variety of sources^1,5,10,11 are that faulty communication is widespread and results in a number of untoward consequences for patients, caregivers, and the organizations in which they are embedded.

A key finding of our study is that failures of communication are not simply the result of faulty transmission and exchange of information. This is not to underestimate the extent to which poor communication results from inadequate information sharing among interdependent caregivers. Certainly our data show that in some relational contexts too little information is communicated, is not timely, or is not communicated using a medium appropriate for the message or task. Yet, our findings show that there is more to it than this mechanistic view of communication suggests. Our findings are consistent with studies in organizational communication^21–25 that show that communication failures arise from vertical hierarchical differences, concerns with upward influence, role conflict and ambiguity, and struggles with interpersonal power and conflict.

Communication is likely to be distorted or withheld in situations where there are hierarchical (e.g., power/status) differences between two communicators,²² particularly when one party is concerned about appearing incompetent, does not want to offend the other, or when one party perceives that the other is not open to communication.²⁵ These situations are most likely to occur in relationships between residents and attendings, medicine residents and other specialties, and also between residents and nurses. Communication difficulties also arise in situations where there are role conflict and ambiguity,^21,23,24 often in the relational context between hospital and community physicians. Finally, the character of communication is sometimes a consequence of personal differences and the interpersonal conflicts that arise from these differences,²⁵ and often occur between residents and nurses.

In sum, barriers to effective communication are both individual (for example communication is impaired when people are busy or fatigued), and systemic. Communication behaviors are embedded in the structure of the organization and reside in the socially structured and culturally patterned behavior of groups (i.e., subunits, specialties, departments) and practices of the institution.^19,21,22,25 Although individuals may decide for themselves how they want to communicate, their behavior is likely to be constrained by the norms of their particular professional subculture.¹⁶ Consequently, remedies must be targeted toward multiple levels: It is going to take more than simply changing one's individual-level actions to make a difference in the system. Actions need to be taken at the level of the group, subunit and organization. A clearer understanding of the underlying dynamics revealed in this study highlight possibilities for future research as well as appropriate interventions both in medical education and in health care organizations themselves.

Our study focused on the inpatient setting. Yet, care is frequently delivered in the outpatient setting. One avenue for future research is to examine whether the factors and dynamics found here are similar to the dynamics that occur in the outpatient setting. A second avenue of future research is to investigate the determinants of caregivers’ communication choices. To our knowledge, little attention in medical education research has been paid to the idea of matching the communication medium to the message. Yet, evidence suggests that the mode of communication often determines the outcomes of a task.^19,26

One recommendation for improving systemic communication comes from studies of adverse events in high risk settings which resemble health care contexts (e.g., aircraft carrier flight operations and wildland firefighting); settings where there is huge variability in circumstances, the need to adapt processes quickly, a quickly changing knowledge base, and highly trained professionals who must use expert judgment in dynamic settings.^1,27 Consistent with the results presented here, other studies have shown that communication failures are important contributors to adverse events in other contexts. One suggestion for improving communication in these circumstances is to enact a five-part briefing protocol (STICC) that is being used by the U.S. Forest Service to give direction to fire fighters:²⁸

• Situation: Here's what I think we face;
• Task: Here's what I think we should do;
• Intent: Here's why;
• Concern: Here's what we should keep our eye on;
• Calibrate: Now talk to me. Tell me if you don't understand, cannot do it, or see something I do not.

Just as clinical practice guidelines can assist practitioners in making decisions and taking actions for specific clinical circumstances, communication practice guidelines like the one noted above can serve the same purpose.

This exploratory study has both strengths and limitations. For example, data from practicing health care providers are a rich and neglected source of information about factors in their own environment that might contribute to error. Because mishaps arise out of a complex interplay of human and organizational factors, a methodology is required that takes into account the actors’ understanding of the event. Self-reports of critical incidents are considered by many, including the aviation industry,²⁹ as the most viable method to obtain comprehensive information about an event, especially when respondents consider a broad range of factors.³⁰ We sought to generate a rich description of residents’ experiences and the contexts in which they are embedded to better understand the basic mechanisms (the “why” and the “how”) of error occurring in everyday organizational settings. We did not intend to establish incidence rates or to establish causality, and patterns that emerged in the data are suggestive of associated causes rather than definitive. We are mindful that retrospective event histories are subject to perceptual biases, poor recall, and poor or inaccurate articulation that could have influenced the results.

In conclusion, faulty communication increasingly is being implicated in the evolution of medical errors and adverse events. And the proposed solution often is to promote “better information transfer.” This goal is hard to achieve for clinicians in the best of circumstances, but it is even more challenging for residents who are caught in a web of complex relationships. Moreover, communication failures entail more than the faulty transfer of information. They are a consequence of individual, relational, and systemic factors, which suggests that more effective communication is more difficult than it looks. We were surprised by the insidiousness of communication failures as contributors to the mishaps reported by our respondents. However, our observations have strong face validity and suggest that the systemic problems of poor communication as latent precursors to medical mishaps should not be taken lightly.

Acknowledgments

This study was supported in part by a grant from the University of Michigan Office of the Vice-President for Research. The authors are grateful to Pat Cornett, Eric Eisenberg, Zach Lewton, Lexa Murphy, Steve Schenkel, Tim Vogus, Bob Wears, and Karl Weick for constructive comments on previous drafts.