R E S E A R C H R E P O R T

Communication Failures: An Insidious Contributor

to Medical Mishaps
Kathleen M. Sutcliffe, PhD, Elizabeth Lewton, PhD, MPH, and Marilynn M. Rosenthal, PhD
Downloaded from https://journals.lww.com/academicmedicine by BhDMf5ePHKav1zEoum1tQfN4a+kJLhEZgbsIHo4XMi0hCywCX1AWnYQp/IlQrHD3XLe684GKHSWEOjO0MoWl0a5+FO3axgHKL5rKcluaavY= on 07/22/2019

ABSTRACT

Purpose. To describe how communication failures con- patient management vis-à-vis other medical staff and
tribute to many medical mishaps. health care providers from within the hospital and from
Method. In late 1999, a sample of 26 residents stratified the community. Recurring patterns of communication
by medical specialty, year of residency, and gender was difficulties occur within these relationships and appear to
randomly selected from a population of 85 residents at a be associated with the occurrence of medical mishaps.
600-bed U.S. teaching hospital. The study design in- Conclusion. The occurrence of everyday medical mis-
volved semistructured face-to-face interviews with the haps in this study is associated with faulty communica-
residents about their routine work environments and tion; but, poor communication is not simply the result of
activities, the medical mishaps in which they recently had poor transmission or exchange of information. Commu-
been involved, and a description of both the individual nication failures are far more complex and relate to
and organizational contributory factors. The themes re- hierarchical differences, concerns with upward influence,
ported here emerged from inductive analyses of the data. conflicting roles and role ambiguity, and interpersonal
Results. Residents reported a total of 70 mishap inci- power and conflict. A clearer understanding of these
dents. Aspects of “communication” and “patient manage- dynamics highlights possibilities for appropriate interven-
ment” were the two most commonly cited contributing tions in medical education and in health care organiza-
factors. Residents described themselves as embedded in a tions aimed at improving patient safety.
complex network of relationships, playing a pivotal role in Acad Med. 2004;79:186 –194.

Medical mishaps are a pervasive prob- studies on which the IOM relied for its organizations do not usually have a sin-
lem in health care organizations. In estimates,2 as well as other studies of gular cause, but result from a string of
2000, the Institute of Medicine (IOM)1 medical error,3,4 were based on physi- latent flaws throughout the system.5,6
reported that between 44,000 and cians’ reviews of medical records and The growing interests in the broader
98,000 people die every year in U.S. their judgments of adverse events. Stud- context of medical mishaps and the sys-
hospitals because of medical errors. The ies in this vein tend to focus on the tems in which health care providers are
incidence and nature of errors in medi- embedded comes at a time when the
Dr. Sutcliffe is associate professor of organizational cine and consider errors a function of conceptual and methodological tools in
behavior and human resource management, Univer- deficiencies in expertise and training. current use drastically underestimate
sity of Michigan Business School, Ann Arbor, Mich- What is missing from current research is the role of social, relational, and orga-
igan. At the time of this study, Dr. Lewton was a
research associate in the University of Michigan Busi- a rich description of the way clinicians nizational factors in the generation of
ness School. Dr. Rosenthal is professor emerita of in complex organizations such as hospi- adverse medical events.5– 8 With these
sociology and currently adjunct professor in internal
medicine, University of Michigan Medical School,
tals experience errors in the context of issues in mind, we undertook this qual-
Ann Arbor, Michigan. daily clinical practice. itative study to examine the individual
Correspondence and requests for reprints should be To be sure, individuals make errors. resident’s experience and perception of
addressed to Dr. Sutcliffe, University of Michigan Yet, studies of organizational accidents the causes and contexts of medical mis-
Business School, 701 Tappan, Ann Arbor, MI
48109-1234; telephone: (734) 764-2312; e-mail: in other disciplines suggest that most of haps. Using data collected through a
具ksutclif@umich.edu典. the accidents that occur in complex series of 26 semistructured interviews

186 ACADEMIC MEDICINE, VOL. 79, NO. 2 / FEBRUARY 2004

with residents regarding their routine large contributor to adverse clinical approach that allowed residents to re-
work environments and the medical events and outcomes. spond to themes generated from the
mishaps with which they had been in- literature and incorporated into the in-
volved, we explored how communica- terview questions, and also enabled
tion plays an integral role in many un- METHOD them to reflect on their own unique
toward events. experiences. Before the interview, re-
Faulty communication has been im- Study Location and Sample spondents were assured of confidential-
plicated both in exhaustively studied ity and anonymity. Each person signed
and well-known catastrophes such as The research site was a 600-bed U.S. and returned a confidentiality and vol-
the explosion of the space shuttle Chal- teaching hospital with a large graduate untary participation agreement and
lenger,6,9 the release of methocyanate in medical education program. Our focus gave us permission to audiotape the in-
Bhopal, India,6 and in the occurrence of was on the inpatient setting. The study terview.
adverse events in other high-risk con- was approved by our university human The interview protocol consisted of a
texts.10,11 Communication failures are subjects’ committee, and the hospital’s series of open-ended questions that fo-
increasingly being implicated as impor- Internal Review Board, Quality com- cused on the general work environment
tant latent factors influencing patient mittee, and department of risk manage- and the medical mishaps in which each
safety in hospitals as well.10,12 A retro- ment. respondent had been involved within
spective Australian survey of hospital The sample, stratified by specialty, the preceding three months. Many
admissions showed that communication year of residency, and gender, was ran- terms have been used to describe medi-
problems were the most common cause domly drawn from a total population of cal errors but we chose to use the term
of preventable disability or death.13 85 residents within the three specialty mishap, a neutral, nonjudgmental term
More recently, Chassin and Becher12 residency programs sponsored by the that encompasses a wide range of mis-
analyzed how it was possible for the hospital: Surgery, Medicine, and Ob- takes from the near-miss to a serious
wrong patient to undergo an invasive stetrics/Gynecology (Ob/Gyn). Al- iatrogenic injury, to elicit a broad range
procedure and concluded that “frighten- though the research site was chosen for of incidents. We asked respondents to
ingly poor communication” was a key convenience, we chose to randomly describe each mishap in depth and sub-
causal element. sample residents to more accurately rep- sequently asked them to categorize each
Today, medical care involves shorter resent the total population. mishap into one or more of six catego-
hospital stays with a rapid turnover of The final sample consisted of 26 res- ries (e.g., omission, commission, diagno-
acutely ill patients. Patient manage- idents (30% of the population) includ- sis, treatment, medication, or patient
ment involves complex investigation ing five surgery, 17 medicine (11 medi- management), which we derived from
and coordination of care by a myriad of cine and six preliminary/transitional), review of existing literature on medical
medical specialists. Clinical medicine and four Ob/Gyn residents. Fourteen mistakes. We then asked each respon-
thus involves multiple handoffs with were men and 12 were women. Eleven dent to describe and categorize the con-
many places where critical information residents were in their first year of post- tributory factors associated with each
must be effectively communicated. In graduate training, five in their second mishap (i.e., elements they thought
addition to the horizontal differentia- year, seven in their third year, two in were linked with the mishap) and to
tion of labor, vertical divisions of hier- their fourth year, and one in their fifth choose the most important factor. We
archy and power operate as well.14 year. The disproportionate number of asked additional questions about resi-
Complex systems are made up of indi- first-year residents in the sample and dents’ work environments and daily activ-
viduals at different hierarchical levels population was the result of students ities, how mishaps were acknowledged,
who must constantly interrelate. These going on to residencies in other pro- and ended with a question about how
relationships are shaped by the relative grams after completing the first year. mishaps could be avoided. Interviews
status of those involved. These struc- Residents ranged in age from 25–39 lasted between one to two hours. All in-
tures can have a powerful influence on years and averaged 29.8 years. terviews were tape-recorded, transcribed,
whether and how critical information is and assigned a unique record number.
effectively communicated. With this Data were analyzed using case analy-
theoretical framework in mind, in this Data Collection and Analysis sis techniques suggested by Yin15 that
study we sought to understand how so- included putting information into dif-
cial, relational, and organizational We contacted individuals first by letter ferent arrays, creating a matrix of cate-
structures contribute to communication and then by telephone. We collected gories and placing evidence within
failures that have been implicated as a data using a semistructured interview these categories, tabulating the fre-

ACADEMIC MEDICINE, VOL. 79, NO. 2 / FEBRUARY 2004 187

quency of different events, and conduct- tem in place for several years. (The total part of the story. Other aspects of faulty
ing an iterative process to build an ex- of all categories exceeds 70 because res- communication repeatedly show up
planation. Transcribed interviews were idents often categorized an incident within the context of four key relation-
entered into a qualitative data computer into more than one category. For exam- ships and provide a framework for un-
program to facilitate coding and sorting ple, a case where a patient with cardiomy- derstanding how issues of miscommuni-
of interview data and to compile simple opathy was sent home from the emer- cation can contribute to many medical
counts and statistics once data had been gency department with a pneumonia mishaps. Representative quotes from
coded. Initial codes were derived from diagnosis was categorized as omission, di- the interviews are presented in Table 1.
the interview questions and the litera- agnosis, and patient management.) These and other quotes are discussed
ture search (see previous paragraph for Practitioner’s knowledge (30) and more fully below.
an example of the six categories for communication (28) were cited most
classifying each mishap), or were in- frequently as the most important factors
duced from the narratives. For example, that contributed to each mishap dis- Relational Dyads and Common
after reviewing all responses to the ques- cussed. On the surface, practitioner’s
Communication Failures
tion “how was the mishap acknowl- knowledge seems to refer to the extent
edged,” we found that the answers could to which a practitioner’s medical
Residents and attendings. Faulty
be sorted into one of five categories: knowledge is complete and accurate; yet
communication typically arises in the
discussion, written documentation, re- further content analyses showed that
context of the relationship between a
view at conference, ignored, or un- this category was more often used to
resident and the attending physician
known. Two researchers coded and refer to a practitioner’s awareness of
with whom the resident works.16 Sev-
sorted interview data into the catego- certain pertinent information. For ex-
eral themes are prominent. The attend-
ries. The two coders agreed in 80% of ample, residents chose this category for
ing is both a supervisor and a teacher,
the cases and all disagreements were information regarding a patient’s test
whereas the resident performs the bulk
resolved through discussion. In addition results, previous diagnoses/treatments,
of patient care and decision making. We
to the categorical data coding, two re- or other medical historical issues that
searchers read through the interview were not communicated effectively to found that residents in our study were
transcripts repeatedly and then dis- the practitioner. Thus, we counted in- concerned about appearing incompe-
cussed them to induce the set of emerg- stances of lack of information as in- tent in front of those with more power
ing themes and patterns and build an stances of faulty communication. We and they were hesitant to communicate
explanation. found other contributory factors that information that was unfavorable or
involved issues of communication/infor- negative to themselves. A resident does
mation transmission as well, including not want to appear ignorant about a
RESULTS
factors related to the specific situation patient. He or she wants to appear
(e.g., busy emergency department), work knowledgeable about the medical con-
Overview
environment (hostile superior), and prac- dition and about the patient in particu-
Respondents reported a total of 70 mis- titioners’ interpersonal skills. In total, lar, and to present pertinent but not
haps, which varied from relatively mi- communication failures of one kind or unnecessary information.
nor incidents such as a near miss in another were an associated or contribu- In our study, the hesitancy to com-
which two patients’ orders were tory factor in 64 mishaps (91%). municate information to superiors was
switched, but subsequently corrected by The narratives themselves reveal ad- evident in situations where a resident
a nurse, to relatively major incidents, ditional insight. In reflecting on their did not want to appear incompetent and
such as a chest tube inserted on the roles within the hospital, residents de- also in situations where residents
wrong side. Outcomes ranged from no scribed themselves as embedded in a thought they may offend those in
untoward consequences to death. Re- complex network of relationships. The power. Residents repeatedly commented
spondents categorized mishaps most fre- resident plays a pivotal role in patient on their hesitancy to call an attending
quently as errors of patient management care and management, vis-à-vis other in the middle of the night. Although it
(29), errors of omission (26), errors of medical staff and administrators within is considered appropriate under certain
diagnosis (24), and treatment (24), and the hospital as well as community med- conditions, unnecessarily waking or
commission (21). Surprisingly, there ical practitioners. Although our qualita- bothering an attending is bound to
were very few medication-related errors tive analysis suggests that poor transmis- cause some anxiety and even some fric-
(6), probably because this institution sion and exchange of information tion. There is a tension between want-
has had a computerized order-entry sys- accounts for some mishaps, this is only ing to be sure one is taking the correct

188 ACADEMIC MEDICINE, VOL. 79, NO. 2 / FEBRUARY 2004

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

Communication Theme

Dyad Hierarchy/Power/Social Structure Lack of Information Mode of Communication/Misinterpretation

Residents/attendings “There has to be free communication “. . .the nurses will not take a patient unless “I like to do everything face to face if I
between the attending physician and the they’ve gotten a report. . .but doctors can. It’s a convoluted system here.
supervising resident and the interns. . .. simply don’t do that. They transfer patients, And I think the more you communicate
You can’t be afraid to or ashamed to ask they do this, they do that and then they the better off you’ll be. You’ll improve
for help.” just don’t tell you.” patient care.”

Residents/community physicians “The lack of communication between the “I was uncomfortable because I didn’t really “. . .the communication between our team
outpatient setting and the office to the know what to do with the patient. I had a and [the community specialist] was
hospital, it is a huge thing. A lot of time patient I didn’t have a whole lot of such that there must have been a
when stuff happens I think it’s always a information on. . .and I had a resident that breakdown there in
lack of communication between was irritated because he had gotten no misunderstanding. . .. You’ve got to get
physicians or a disagreement about communication from the admitting the information through all those
what’s the correct course of action.” physician.” points. . ..”

Medicine Residents/specialists “. . .it’s wasting your time dealing with that “When a patient comes from the ER to the “So the information that the ER gave to
because [the emergency room (ER)] will floor the nurse always gets a complete me at 10:30 is now 12 hours old;
deny anything happened. . .. It’s basically report. But most of the time it’s doctors three additional studies have been
a futile attempt. . .. You’re not going to who are not very good about it. And obtained. All the results have been
accomplish anything. . ..” anything can happen and nobody ever phoned to a different person; it’s just a
knows. It’s ridiculous.” pain. . ., just like the telephone game
where you rehash the story time and
time again.”

Residents/nurses “[The nurse] felt that I was not adequately “I think my lack of communication with the “It’s easier to communicate with the
addressing her concerns about the student nurse [contributed to the mishap]. nurses on a person-to-person basis
patient. . .. She said she felt intimidated I think if it had been a more experienced rather than just putting orders in. But
by me.” nurse she wouldn’t have moved the patient. if you’re not there and you put labs
However I could have communicated better through in the morning. . ., it’s not
with her. I should have said something like done unless you go up and ask them
’don’t touch anything. I’ll be right back.’” what’s going on.”

action and wanting to know enough not described why she did not openly object tensive care, and survived with serious
to have to contact the attending. to the treatment: complications. Although it is not clear
Residents’ concerns about offending which course of action was best in this
those in power combined with their If I felt like I could actually communi- case, closed lines of communication be-
perceptions that powerful others would cate with that group of attendings I tween the attending and the resident
not listen to them or hear their point of would have tried, but I didn’t feel like it may have impeded optimal patient care.
view also discouraged residents in our would be useful for me. And all it would Another kind of miscommunication
study from productively disagreeing have done would be to inflame the rela- between attending physicians and resi-
when they had a different point of view. tions between me and that attending and dents concerns the amount of infor-
In one case, the attending physician the patient still would have ended up mation communicated by superiors.
treated an elderly woman with a super- getting [inappropriate treatment]. Specifically, residents perceived that at-
ficial vein thrombosis with anticoagu- tendings provided far too little informa-
lants. The resident vehemently dis- The patient had a massive retroperi- tion to the resident who would be caring
agreed with this course of action, and toneal bleed, was transferred to the in- for a particular patient.

ACADEMIC MEDICINE, VOL. 79, NO. 2 / FEBRUARY 2004 189

 As one resident said: who are ultimately responsible for tak- the intern didn’t misunderstand the
ing care of the patient in the hospital whole thing and maybe I didn’t explain
The nurses will not take a patient un- right then. it enough to her, and again it’s that
less they’ve gotten a report. And they whole A to B to C to D. You’ve got to
are very strict about that. And you Communication failures within this get the information through all those
know, doctors simply don’t do that. relationship also arise out of role con- points and there’s a break down. I
They transfer patients, they do this, flict and ambiguity. Occasionally, a pri- don’t know where it happened neces-
they do that, and then they just don’t sarily. . . . It’s stressful on the residents
vate community physician admits a pa-
tell you. They don’t tell the resident. and the interns because we’re caught in
And, I mean, it’s a simple thing but,
tient to the academic team but then
the middle. We can’t make our own
you know, a five-minute discussion maintains some role in the patient’s decisions, and yet the specialists come
about a patient that’s being admitted care. In this context, disagreement along whenever they want. So we’re
or being transferred or whatever is about patient care and management, as either bugging them, you know, calling
worth its weight in gold when some- well as miscommunication, are com- them, or there’s a break down in com-
thing dreadful happens and I’m run- mon. Friction may occur when the res- munication because it’s going from the
ning down there and I don’t even ident and the community physician dis- specialists usually to the intern to the
know how old the patient is or what resident and the attending. It’s like [the
agree on patient care decisions, or
problems they have. They really need game] telephone, you know?
miscommunicate about patient manage-
to communicate better.
ment.17,18 In one case related by a resi-
dent, a 30-year-old woman was admit- This case illustrates two communica-
This type of situation occurred be-
ted by her private physician on a Friday tion issues. First, little information was
tween residents and attending physi-
morning and was scheduled for mag- communicated to the resident who was
cians within the hospital, and also be-
netic resonance imaging (MRI) that af- asked to carry out the plan of care (the
tween residents and consulting
ternoon. Her physician suspected a dis- MRI). Second, the communication
physicians from within the hospital and
section in the carotid artery. The links were so convoluted that any mes-
from the larger community.
Hospital residents and community resident was aware of the MRI, but had sages that were conveyed were misinter-
physicians. Lack of information and/or no other information about the patient preted as they moved through all those
the ineffective communication of infor- and did not suspect any particular diagno- involved. In sum, insufficient informa-
mation arising from the relationship be- sis. The resident sent the patient home for tion, faulty exchanges of existing infor-
tween hospital residents and commu- the weekend pending the MRI results and mation, or ambiguous and unclear infor-
nity physicians were associated with was severely chastised by the private phy- mation seem to characterize incidents
reported mishaps. This faulty communi- sician for doing so against his wishes. The involving hospital residents and com-
cation occurred most often when the MRI results, which were discovered Mon- munity physicians. The poor communi-
hospital resident admitted a patient day morning, confirmed dissection in the cation often arises out of role conflict
who was previously under a community carotid artery—a potentially serious con- and ambiguity—where the boundaries
physician’s care. At the hospital in dition. The resident highlighted some ar- between who has the authority for care
which we gathered our data, private eas of miscommunication: and who is responsible for care are un-
physicians may admit a patient whose clear. Our evidence suggests that these
care is then taken over by the “academic Our suspicion of the pathology wasn’t situations provide fertile ground for
team” of staff physicians and residents. there because [the patient looked mishaps.
During this complete transfer of patient healthy and we didn’t know what her Internal medicine residents and spe-
care, information about the patient is physician had in mind]. . . . So we cialists. Faulty communication is perva-
often seriously or completely lacking. asked him [her physician] if we could sive in relationships between medicine
As one resident described: just get the MRI and then discharge residents and the residents and/or at-
her for him to follow-up on the results. tendings in different specialties within
God, we get patients, we’re lucky if we And again, this is me communicating
the hospital. Our analysis revealed some
get any information when the patients through my intern and then my intern
reasons for this. Not surprisingly, the
come into the hospital. About half the talking to him. And it was fine, we got
the MRI and let her go. The commu- timely and effective exchange of perti-
time we don’t even know they’re com-
ing. They’ll just show up on the floor. nication between our team and, you nent information was a prominent
There’s a huge, huge potential for mis- know, the specialist was such that theme, particularly between the emer-
haps to occur that way . . . between the there must have been a break down in gency department and internal medi-
admitting physician and the people understanding. And I’m wondering if cine, as one medicine resident noted:

190 ACADEMIC MEDICINE, VOL. 79, NO. 2 / FEBRUARY 2004

 Once they called me from the ER on a We were lucky that we got to the struggles often led to patient manage-
patient who came with congestive patient soon enough because had the ment problems. In one case, a cardiolo-
heart failure. They didn’t mention that patient been sitting there, as it happens gist who was called to consult on a
the patient had EKG changes and ab- at times, depending on how busy the patient put the patient on a medication.
normal heart enzymes. She was having intern and resident are, and not [been]
The medicine attending took her off the
a heart attack but we were really super assessed by anyone, I think we could
have had a bad outcome.
medication. The notes in the patient’s
super busy so we didn’t have a chance
record were unclear and there was no
to take a look at the patient until she
We noted earlier how the openness verbal communication between the par-
hit the floor and she was having a heart
and quality of communication suffers in ties. This led to a disagreement with
attack. She was infarcting. And so
that’s communication. Communica- the presence of private concerns about regard to patient management and the
tion’s the problem. When somebody’s hierarchy and power. Communication is resident felt stuck in the middle:
taking care of the patient, when the distorted because of concerns about of-
patient’s transferred from a service to a fending the more powerful party, and . . . I guess the attending that was on
different service, there should be very also because residents want to avoid our team had a different philosophy
good communication for that so that conflict or believe that their concern than the cardiologist who would have
kept her on anticoagulants. And, . . .
all the information is conveyed there. will not be addressed. These issues sur-
you know, in his note [he] didn’t write
face frequently in situations where resi-
it down and never verbalized it to us. I
In another case, a resident reported dents interact with the attendings in mean, just bad communication.
her experience with a patient who had other departments, which is often re-
just been admitted and transferred to quired of them. For example, medicine As this resident went on to explain,
residents frequently admit patients sent much of the communication between
general medicine from the emergency
to them by emergency department at- consultants and medicine attendings is
department with dangerously low blood
tendings. When problems arise, a resi- through written notes in the patient’s
pressure. No one on the floor had been
dent may find it difficult and frustrating record—a sometimes ineffective means
notified of the patient’s condition be-
to speak up, particularly to disagree, of communicating:
fore arrival. The resident to whom the with an unknown attending. Residents
patient was assigned just happened to feel uncomfortable in voicing their objec-
walk by the room when the patient The charts are the primary means of
tions and perceive that even if they do, communication between a lot of differ-
started to decompensate: they may not get a positive response. One ent people—the consultants, the in-
resident in our study described what hap- terns and everyone. It’s not necessarily
As I walked in, the patient was in pened when a resident disagreed with the the best way because a lot of times you
relative distress, short of breath, looked patient management prescribed by the can’t read the handwriting very well. A
sick. And so I asked them to take a emergency department: lot of times something’s forgotten,
blood pressure immediately and the you’re just writing a quick note and
blood pressure was in the seventies It’s kind of hard for them to accept they forget to mention something.
over doppler. So I immediately put the [disagreement] because they are at-
patient in Trendelenburg and got some tendings and we are residents—a lower Residents and nurses. The relation-
IV fluids going and the patient looked level of the hierarchy. So it’s not that ship between residents and nurses is
septic to me. And I had to send the nice—they don’t feel that good if they critical in preventing medical mishaps.
patient to the unit immediately. I get a call from the resident and they
Residents make most of the day-to-day
think that patient should have never are an attending and that “blah, blah,
blah happened and you guys misman-
decisions about patient care and nurses
come to the [medicine] floor to begin not only carry out many of the orders,
aged that patient.”
with. And when I look back, the nurs- but also are in closest contact with the
ing records show[ed] that the patient patient and are better informed about
There is evidence that role conflict
had been hypotensive in the ER prior
and contrasting core values between dif- their moment-by-moment condition.
to even coming up to the floor.
ferent kinds of physicians can influence Information given to nurses by residents
effective communication.10,17 These in the form of orders, and information
Fortunately, the outcome in this case tendencies are exacerbated by the fact given to residents by nurses in the form
was good. The patient was moved to the that very little direct (face-to-face) of the patient’s condition are crucial. Our
intensive care unit soon enough and communication occurs between primary analysis revealed why both kinds of in-
avoided any serious complications, as care physicians and specialist consult- formation may not be communicated
the resident noted: ants.18 In our analysis, these kinds of effectively.

ACADEMIC MEDICINE, VOL. 79, NO. 2 / FEBRUARY 2004 191

 Residents routinely communicate or- staff that “We appreciate you coming There was a communication problem
ders in hospitals by writing or typing the down and telling us this so that it between the nurse who was taking care
orders in the patient’s medical record. doesn’t sit around for three hours and of her and myself in that she felt that I
Yet, written communication, an imper- the patient doesn’t get their nitroglyc- was not adequately addressing her con-
erine that they need.” Or whatever. cerns about the patient, so she didn’t
sonal medium with limited capacity for
call me back to say that the patient was
timely feedback,19 is seldom the most continuing not to do well . . . I don’t
The personal characteristics of com-
effective way to communicate a plan of understand. She said she felt intimi-
municators also can influence the char-
action, especially when action needs to dated by me or something. It really didn’t
acter of a communication relationship,
be taken quickly. As one resident noted: make a lot of sense to me at the time. . . .
sometimes giving rise to interpersonal
We actually sat down and had a discus-
I think the main important thing is conflict.19,20 For example, in one case sion about this after—with the attending
communication with the nurses. Put- the resident made a mistake in writing a physician and the nurse and myself—
ting orders in the computer or just blood pressure medication prescription about how this had transpired and what
putting orders in the chart can be for his patient. The nurse noticed this had happened. And so that’s how I had
missed and they have been missed be- and mentioned it right away, but not in some insight into her perspective on it.
fore just because people don’t look at such a way that the resident understood And basically she felt that . . . I was
it. And if you don’t tell the nurse on and/or responded. The nurse mentioned blowing her off or didn’t seem to be
the floor that “this is going to be or- it again only after the incorrect dosage approachable and therefore for some rea-
dered, watch for it,” like that, it may was administered: son she felt it was appropriate just not to
take them a couple hours to go and get do anything until the next day.
the order from the clerk and do the
appropriate work. And I’ve seen it The nurse had questions and did bring
it to my attention, but after the medi- Although the surgery resident was
more than not that that happens—that not certain of the eventual outcome
orders get missed. And just recently cine had been given. Now she said that
she had asked me as well before she because the patient was transferred to
there was a chemo treatment order
that was missed. It was in the chart gave it. But she certainly asked me in a another unit, he reported that there
early in the morning and it was missed much different way after she gave it. So were “potential long-term complica-
until three o’clock in the afternoon. she could have come to me and been a tions” for the patient due to this mishap.
And I was asking, “Where’s the treat- little bit more forceful about how it In the following excerpt from an in-
ment?” and they looked in the chart didn’t make sense. terview, a resident explained the diffi-
and said, “Oh nothing was started” cult position nurses often find them-
because the attending never flagged The nonverbal information displayed selves in.
the chart or didn’t tell any of the by one or the other party (communi-
nurses, so no one looked at it. And cated through eye gaze, posture, facial Resident: In some cases nurses can
then when I was called by the attend- expression, and voice tone) can also be reluctant to question physicians
ing in the afternoon [to check] that the influence communication,21 and our because, you know, because of physi-
chemo started, I asked the nurses and data reflect this. For example, several cians’ responses to them, questioning
they said no. But there was no nurse to residents suggested that whether the them, or residents for that matter.
run the chemo at night. So it got delayed Primary investigator: You mean
nurse was “nice” or “aggressive” could
until the next morning. That could have they have a sense that they [physi-
been avoided by communication.
make a difference in how an intern
responded to a nurse’s concerns regard- cians] don’t welcome criticism, any
ing a patient. questioning?
Here the delay in treatment occurred Resident: Right. So I think this
without adverse effects, but this type of Although personal characteristics
demonstrates that it’s very important
miscommunication could result in more may matter, we repeatedly found that
that nurses feel able to question the
serious consequences. communication behaviors are influ- doctors...they walk a fine line because
Another resident gave a similar opin- enced by hierarchy and social structure they don’t always know what’s appro-
and, as described below, perceptions priate to question and what’s not ap-
ion and also noted the importance of
that a superior is receptive to receiving propriate to question. . . .
face-to-face interactions:
information. For example, as one resi-
I think that communication, especially dent related, a 70-year-old-patient who
between the doctors and the nurses had recently had cardiac bypass surgery DISCUSSION
who are giving the stuff that you’re was not urinating and had low blood
ordering, is very key. And I’ve had a lot pressure during the night following Our qualitative study provides insight
of positive feedback from the nursing surgery: into the insidiousness of faulty commu-

192 ACADEMIC MEDICINE, VOL. 79, NO. 2 / FEBRUARY 2004

nication as a contributor to medical idents and nurses. Communication dif- dium to the message. Yet, evidence sug-
mishaps. The residents in our study per- ficulties also arise in situations where gests that the mode of communication
ceived that communication difficulties there are role conflict and ambigu- often determines the outcomes of a
played a role in the vast majority of the ity,21,23,24 often in the relational con- task.19,26
medical mishaps they experienced. Our text between hospital and community One recommendation for improving
findings are consistent with other re- physicians. Finally, the character of systemic communication comes from
search showing a strong link between communication is sometimes a conse- studies of adverse events in high risk
poor communication, and errors and ad- quence of personal differences and the settings which resemble health care
verse events.12,13 Although few studies interpersonal conflicts that arise from contexts (e.g., aircraft carrier flight op-
in medicine have systematically exam- these differences,25 and often occur be- erations and wildland firefighting); set-
ined this association, the implications tween residents and nurses. tings where there is huge variability in
drawn from a variety of sources1,5,10,11 In sum, barriers to effective commu- circumstances, the need to adapt pro-
are that faulty communication is wide- nication are both individual (for exam- cesses quickly, a quickly changing
spread and results in a number of unto- ple communication is impaired when knowledge base, and highly trained pro-
ward consequences for patients, caregiv- people are busy or fatigued), and sys- fessionals who must use expert judg-
ers, and the organizations in which they temic. Communication behaviors are ment in dynamic settings.1,27 Consis-
are embedded. embedded in the structure of the orga- tent with the results presented here,
A key finding of our study is that nization and reside in the socially struc- other studies have shown that commu-
failures of communication are not sim- tured and culturally patterned behavior nication failures are important contrib-
ply the result of faulty transmission and of groups (i.e., subunits, specialties, de- utors to adverse events in other con-
exchange of information. This is not to partments) and practices of the institu- texts. One suggestion for improving
underestimate the extent to which poor tion.19,21,22,25 Although individuals may communication in these circumstances
communication results from inadequate decide for themselves how they want to is to enact a five-part briefing protocol
information sharing among interdepen- communicate, their behavior is likely to (STICC) that is being used by the U.S.
dent caregivers. Certainly our data show be constrained by the norms of their Forest Service to give direction to fire
that in some relational contexts too particular professional subculture.16 fighters:28
little information is communicated, is Consequently, remedies must be tar-
not timely, or is not communicated us- geted toward multiple levels: It is going 䡲 Situation: Here’s what I think we
ing a medium appropriate for the mes- to take more than simply changing face;
sage or task. Yet, our findings show that one’s individual-level actions to make a 䡲 Task: Here’s what I think we should
there is more to it than this mechanistic difference in the system. Actions need do;
view of communication suggests. Our to be taken at the level of the group, 䡲 Intent: Here’s why;
findings are consistent with studies in subunit and organization. A clearer un- 䡲 Concern: Here’s what we should keep
organizational communication21–25 that derstanding of the underlying dynamics our eye on;
show that communication failures arise revealed in this study highlight possibil- 䡲 Calibrate: Now talk to me. Tell me if
from vertical hierarchical differences, ities for future research as well as appro- you don’t understand, cannot do it, or
concerns with upward influence, role priate interventions both in medical ed- see something I do not.
conflict and ambiguity, and struggles ucation and in health care organizations
with interpersonal power and conflict. themselves. Just as clinical practice guidelines can
Communication is likely to be dis- Our study focused on the inpatient assist practitioners in making decisions
torted or withheld in situations where setting. Yet, care is frequently delivered and taking actions for specific clinical
there are hierarchical (e.g., power/sta- in the outpatient setting. One avenue circumstances, communication practice
tus) differences between two communi- for future research is to examine guidelines like the one noted above can
cators,22 particularly when one party is whether the factors and dynamics found serve the same purpose.
concerned about appearing incompe- here are similar to the dynamics that This exploratory study has both
tent, does not want to offend the other, occur in the outpatient setting. A sec- strengths and limitations. For example,
or when one party perceives that the ond avenue of future research is to in- data from practicing health care provid-
other is not open to communication.25 vestigate the determinants of caregivers’ ers are a rich and neglected source of
These situations are most likely to occur communication choices. To our knowl- information about factors in their own
in relationships between residents and edge, little attention in medical educa- environment that might contribute to
attendings, medicine residents and tion research has been paid to the idea error. Because mishaps arise out of a
other specialties, and also between res- of matching the communication me- complex interplay of human and orga-

ACADEMIC MEDICINE, VOL. 79, NO. 2 / FEBRUARY 2004 193

nizational factors, a methodology is re- President for Research. The authors are grateful to 16. McCue JD, Beach KJ. Communication barri-
quired that takes into account the ac- Pat Cornett, Eric Eisenberg, Zach Lewton, Lexa ers between attending physicians and resi-
Murphy, Steve Schenkel, Tim Vogus, Bob Wears, dents. J Gen Intern Med. 1994;9:158 – 61.
tors’ understanding of the event. Self- and Karl Weick for constructive comments on 17. Epstein RM. Communication between pri-
reports of critical incidents are previous drafts. mary care physicians and consultants. Arch
considered by many, including the avi- Fam Med. 1995;4:403–9.
ation industry,29 as the most viable 18. McPhee SJ, Lo B, Saika GY, Meltzer R. How
REFERENCES good is communication between primary care
method to obtain comprehensive infor-
physicians and subspecialty consultants? Arch
mation about an event, especially when 1. Kohn LT, Corrigan JM, Donaldson MS (eds). Intern Med. 1984;144:1265– 68.
respondents consider a broad range of To Err Is Human: Building a Safer Health Sys- 19. Sitkin SB, Sutcliffe KM, Barrios-Choplin JR.
factors.30 We sought to generate a rich tem. Committee on Quality of Health Care in A dual-capacity model of communication me-
description of residents’ experiences and America, Institute of Medicine. Washington, dia choice in organizations. Hum Commun
the contexts in which they are embed- DC: National Academy Press, 2000. Res. 1992;18:563–98.
2. Thomas EJ, Studdert DM, Newhouse JP, et al. 20. Duldt BW. Anger: an alienating communica-
ded to better understand the basic Costs of medical injuries in Colorado and tion hazard for nurses. Nurs Outlook. 1981;
mechanisms (the “why” and the “how”) Utah in 1992. Inquiry. 1999;36:255– 64. 29:640 – 4.
of error occurring in everyday organiza- 3. Hayward RA, Hofer TP. Estimating hospital 21. Dansereau F, Markham SE. Superior-
tional settings. We did not intend to deaths due to medical errors: preventability is subordinate communication: multiple levels
establish incidence rates or to establish in the eye of the reviewer. JAMA. 2001;286: of analysis. In: Jablin FM, Putnam LL, Roberts
415–20. KH, Porter LW (eds). Handbook of Organi-
causality, and patterns that emerged in 4. Thomas EJ, Studdert DM, Brennan TA. The zational Communication. Newbury Park, Cal-
the data are suggestive of associated reliability of medical record review for esti- if.: Sage, 1987;343– 88.
causes rather than definitive. We are mating adverse event rates. Ann Intern Med. 22. Frost PJ. Power, politics, and influence. In:
mindful that retrospective event histo- 2002;136:812– 6. Jablin FM, Putnam LL, Roberts KH, Porter
ries are subject to perceptual biases, 5. Reason J. Managing the risks of organizational LW (eds). Handbook of Organizational Com-
accidents. Aldershot, UK: Ashgate, 1997. munication. Newbury Park, Calif.: Sage,
poor recall, and poor or inaccurate ar- 6. Reason J. Human error: models and manage- 1987;503– 48.
ticulation that could have influenced ment. BMJ. 2000;320:768 –70. 23. Jablin FM. Task/work relationships: A life-span
the results. 7. Taylor-Adams S, Vincent C. Clinical acci- perspective. In: Knapp ML, Miller GR (eds).
In conclusion, faulty communication dent analysis: understanding the interactions Handbook of Interpersonal Communication.
increasingly is being implicated in the between the task, individual, team and orga- Newbury Park, Calif.: Sage, 1985;615–54.
nization. In: Vincent C, de Mol B (eds). 24. Jablin FM. Formal organization structure. In:
evolution of medical errors and adverse Safety in Medicine. Amsterdam: Pergamon, Jablin FM, Putnam LL, Roberts KH, Porter
events. And the proposed solution often 2000;101–16. LW (eds). Handbook of Organizational Com-
is to promote “better information trans- 8. Weick KE. The reduction of medical errors munication. Newbury Park, Calif.: Sage,
fer.” This goal is hard to achieve for through mindful interdependence. In: 1987;389 – 420.
clinicians in the best of circumstances, Rosenthal MM, Sutcliffe KM (eds). Medical 25. Stohl C, Redding WC. Messages and message
Error: What Do We Know? What Do We Do? exchange processes. In: Jablin FM, Putnam
but it is even more challenging for res- San Francisco: Jossey-Bass, 2002;177–99. LL, Roberts KH, Porter LW (eds). Handbook
idents who are caught in a web of com- 9. Vaughan D. The Challenger Launch Decision. of Organizational Communication. Newbury
plex relationships. Moreover, communi- Chicago: University of Chicago Press, 1996. Park, Calif.: Sage, 1987;451–502.
cation failures entail more than the 10. Helmreich RL, Merritt AD. Culture at work 26. West DW, Levine S, Magram G, et al.
faulty transfer of information. They are in aviation and medicine: national, organiza- Pediatric medication order error rates related
tional, and professional influences. Aldershot, to the mode of order transmission. Arch Pe-
a consequence of individual, relational, UK: Ashgate, 1998. diatr Adolesc Med. 1994;148(12):1322– 6.
and systemic factors, which suggests 11. Weick KE, Sutcliffe KM. Managing the unex- 27. Weick KE. The collapse of sensemaking in
that more effective communication is pected: assuring high performance in an age of organizations: the Mann Gulch disaster. Adm
more difficult than it looks. We were complexity. San Francisco: Jossey-Bass, 2001. Sci Q. 1993;38:628 –52.
surprised by the insidiousness of com- 12. Chassin MR, Becher EC. The wrong patient. 28. Weick KE. Puzzles in organization learning:
Ann Intern Med. 2002;136:826 –33. an exercise in disciplined imagination. Br J
munication failures as contributors to 13. Wilson RM, Runciman WB, Gibberd RW, et Manage. 2002;13:S7–S17.
the mishaps reported by our respon- al. The quality in Australian health care 29. Reynard WD, Billings CE, Cheaney ES, Hardy
dents. However, our observations have study. Med J Aust. 1995;63:458 –71. R. The development of the NASA aviation
strong face validity and suggest that the 14. Engestrom Y. Developmental studies of work safety reporting system. NASA Reference Publ
systemic problems of poor communica- as a test bench for activity theory: the case of No. 1114. Washington, DC: National Aero-
primary care in medical practice. In: Chaiklin S, nautics and Space Administration Scientific
tion as latent precursors to medical mis- Lave J (eds). Understanding Practice: Perspec- and Technical Information Branch, 1986.
haps should not be taken lightly. tives on Activity and Context. Cambridge: 30. Bogner MS. A systems approach to medical
Cambridge University Press, 1993:64–104. error. In: Vincent C, de Mol B (eds). Safety in
This study was supported in part by a grant from 15. Yin RK. Case Study Research: Design and Medicine. Amsterdam: Pergamon, 2000;83–
the University of Michigan Office of the Vice- Methods. Thousand Oaks, Calif.: Sage, 1994. 100.

194 ACADEMIC MEDICINE, VOL. 79, NO. 2 / FEBRUARY 2004