- 1Lisa I. Iezzoni is a professor of medicine at Harvard Medical School and director of the Mongan Institute for Health Policy at Massachusetts General Hospital, in Boston.
- 2Sowmya R. Rao is an associate professor in the Quantitative Health Sciences Department at the University of Massachusetts Medical School, in Worcester.
- 3Catherine M. DesRoches is a senior researcher at Mathematica Policy Research, in Cambridge, Massachusetts.
- 4Christine Vogeli is a health services researcher at the Mongan Institute for Health Policy.
- 5Eric G. Campbell is an associate professor of medicine at Harvard Medical School and director of research at the Mongan Institute for Health Policy.
- ↵*Corresponding author
The Charter on Medical Professionalism, endorsed by more than 100 professional groups worldwide and the US Accreditation Council for Graduate Medical Education, requires openness and honesty in physicians’ communication with patients. We present data from a 2009 survey of 1,891 practicing physicians nationwide assessing how widely physicians endorse and follow these principles in communicating with patients. The vast majority of physicians completely agreed that physicians should fully inform patients about the risks and benefits of interventions and should never disclose confidential information to unauthorized persons. Overall, approximately one-third of physicians did not completely agree with disclosing serious medical errors to patients, almost one-fifth did not completely agree that physicians should never tell a patient something untrue, and nearly two-fifths did not completely agree that they should disclose their financial relationships with drug and device companies to patients. Just over one-tenth said they had told patients something untrue in the previous year. Our findings raise concerns that some patients might not receive complete and accurate information from their physicians, and doubts about whether patient-centered care is broadly possible without more widespread physician endorsement of the core communication principles of openness and honesty with patients.
In its landmark 2001 Crossing the Quality Chasm report, the Institute of Medicine listed patient-centeredness—care that respects patients’ preferences, needs, and values—as one of six essential features of a high-quality health care system.1 In his 2002 “user’s guide” to that report, Donald Berwick called patient-centeredness the “true north” that should guide all health reform efforts.2
In his “confessions of an extremist,” Berwick expanded on this position.3 The Institute of Medicine committee’s deliberations that ultimately led to patient-centeredness, he explained, were a compromise position between two extremes: the position stated by the late sociologist Eliot Freidson in his early conceptualization of medical professionalism (“Trust us: we know best what will help you”) and the consumerist position (“Let us know what you need and want, and that is what we will offer”).3(p w3)
In arguing that patient-centeredness should be the central, distinct, and dominant dimension of health care quality, regardless of its relationship to other dimensions, such as safety and efficiency, Berwick described three maxims that guided his understanding: “The needs of the patient come first”; “Nothing about me without me”; and “Every patient is the only patient.”3(p w560)
Open, transparent, and bidirectional communication between patients and physicians is the foundation of patient-centered care. Fundamentally, communication by both patients and physicians represents “transmission of information, thoughts, and feelings so that they are satisfactorily received or understood.”4
In most clinical interactions, patients describe symptoms and signs to inform clinical evaluations. After synthesizing this information, physicians indicate diagnoses and suggest treatment plans. Beyond that, communication shapes interpersonal relationships between patients and physicians through mutual trust and empathy, understanding of patients’ preferences and values, and ultimately the course of care.5–11
Movements toward collaborative care, empathic listening, shared decision making, and culturally competent care reflect a growing understanding of the potentially therapeutic implications of patient-physician communication.7,12⇓⇓–15 In 1999 the Accreditation Council for Graduate Medical Education described “interpersonal and communication skills” as skills “that result in effective information exchange and teaming with patients.”16 The council listed those skills among six core competencies that residents must attain during their postgraduate training.
In recent years, the subject of communication between patients and physicians has broadened to encompass more than discussions of patients’ immediate clinical concerns. Specifically, it now includes information that physicians are ethically obligated to convey to patients and that could be essential to patient-centered care.
For instance, recognition of the high frequency of medical errors, the morbidity and mortality caused by those errors, and the resulting risks of malpractice litigation has focused attention on how and when physicians should discuss errors in their care with patients.8,17⇓–19 Another area of concern is how open and honest physicians are in discussing clinical prognoses and the various risks and benefits of interventions with patients.20⇓⇓–23
Other potentially important topics that merit open communication between physicians and patients include relationships between physicians and industry, reimbursement incentives, physicians’ investments in medical facilities, and general concerns about conflicts of interest.21,24,25 Maintaining the confidentiality of patient information is also an issue: The Health Insurance Portability and Accountability Act of 1996 made this both a professional and a legal obligation.26,27
The Charter on Medical Professionalism, endorsed by more than 100 professional groups worldwide and by the Accreditation Council for Graduate Medical Education, addresses communication topics among its three guiding principles and ten responsibilities or commitments.28,29 The charter’s first principle, patient welfare, notes that altruism contributes to the trust that is essential to physician-patient relationships, and it warns against threats from market forces to this trust.
The second principle, patient autonomy, requires physicians to be honest with patients so as to foster informed decision making. Honesty must include promptly informing patients of medical errors and is one of ten professional responsibilities. Another responsibility entails maintaining appropriate relationships with patients. The charter also exhorts physicians not to exploit patients for financial gain, instructing physicians to manage conflicts of interest to retain patient trust.28,29
Despite endorsements by numerous professional associations, it is unknown how widely the charter’s communication provisions are accepted or followed by US physicians. In 2009 we surveyed practicing physicians nationwide to explore their attitudes and behavior regarding different aspects of the charter. This article addresses physician-patient communication concerns and related issues that affect patient-centered care.
Survey Development And Testing
For this study, we revised the professionalism questionnaire that we used in a 2004 national physician survey.30⇓⇓–33 That survey included batteries of questions addressing topics concerning professionalism from each of the charter’s different domains. For the 2009 survey, we revised the 2004 questions that had not adequately discriminated among respondents.34,35
To identify potential new topics for the 2009 survey, we convened a fifteen-member, multispecialty advisory group. We also conducted four focus groups with a total of forty physicians from diverse specialties. The focus-group moderator addressed five broad topics, including honesty with patients. We pretested the revised questionnaire with twenty-one practicing physicians.
The final questionnaire (for details, see the online Appendix36) included five questions about attitudes and four questions about behavior related to physician-patient communication. Because we needed parsimonious question batteries to cover every charter domain without overburdening respondents, we could not ask about all aspects of physicians’ attitudes and behavior related to communication. Instead, we included nine questions as tracer measures of attitudes and behavior.
Institutional Review Boards at the participating institutions approved the questionnaire and administration procedures. Other articles report findings from this project about different charter domains, such as managing conflicts of interest and professional self-regulation.34,35
Sample And Survey Administration
Using the American Medical Association’s 2008 Masterfile, we identified all US physicians in primary care (internal medicine, family practice, and pediatrics) and four other specialties (cardiology, general surgery, psychiatry, and anesthesiology). We excluded doctors of osteopathy, resident physicians, practitioners in federally owned hospitals, retirees, physicians without addresses, and those requesting no contacts. From this pool of eligible participants, we randomly selected 500 physicians within each of the seven specialties, for a total of 3,500 physicians.
The Center for Survey Research at the University of Massachusetts, Boston, administered the eight-page survey in May 2009. The center sent sampled physicians the survey packet, which contained a cover letter; fact sheet; questionnaire, with a sticker on the back containing a random subject identification number; postage-paid return envelope; and a $20 incentive. The center telephoned nonrespondents to solicit participation.
Of the 3,500 sampled physicians, 562 were ineligible because they were deceased, out of the country, practicing a nonsampled specialty, on leave, or not actively practicing. Of the remaining 2,938 eligible physicians, 1,891 completed the survey (a 64.4 percent overall response rate). Response rates varied by specialty, as follows: pediatrics, 72.7 percent; family practice, 67.5 percent; surgery, 65.1 percent; anesthesiology, 64.6 percent; psychiatry, 64.0 percent; internal medicine, 60.8 percent; and cardiology, 50.6 percent.
For the results reported here, we eliminated sixty-one respondents who did not supply usable ZIP code information, which was required to determine regional malpractice claim rates. These sixty-one respondents had demographic characteristics similar to those of the included respondents.
We created binary dependent variables from responses to the five attitude and four behavior questions related to physician-patient communication. We used responses at the highest or lowest ends of the response-category continuum, depending on which end best represented charter precepts.
We focus here on extremes of the response-category continuum for two reasons. First, each question addressed topics that the charter views as key components of professionalism—items representing core professional norms (such as attitudes that physicians should support completely, not partially). Second, responses were not distributed evenly across the categories but were clustered at the extremes that represent endorsement of charter precepts.
As independent (predictor) variables, we included respondents’ sex; underrepresented minority status (race or ethnicity other than white or Asian); years in practice (fewer than 10, 10–19, 20–29, or 30 or more); graduation from a medical school outside the United States or Canada; specialty; and practice setting (university or medical school, hospital or clinic, group practice, solo or two-person practice, and other). Our previous research had found associations between these factors and medical professionalism.30,32,33 We hypothesized that these factors independently affect physicians’ communication attitudes and behavior.
Given a potential relationship between patient-physician communication and malpractice claims,17⇓–19,37 we also hypothesized that local malpractice environments affect physicians’ attitudes and behavior. To create a proxy for this factor, we calculated total malpractice claims30⇓⇓–33 paid per physician per state from the 2009 National Practitioner Database, grouping these numbers into three groups (low, medium, and high).38
All analyses used weights to account for differential sampling rates and nonresponse by specialty. We first examined bivariable associations between dependent and independent variables using two-sided t tests for continuous variables or chi-square tests for categorical variables to test for statistical significance. We used multivariable logistic regression models with associated standard errors to predict our binary-dependent variables, adjusting for the independent variables listed above.39 Analyses used the following statistical analysis software: SAS, version 9.2; and SUDAAN, version 10.0.1.
Our study has several limitations. First, because of social desirability bias, or the tendency of respondents to answer questions in a way that others will view favorably, our results probably represent a lower-bound estimate of the actual frequency of communication failures. Second, although we achieved an excellent response rate for a physician survey, nonresponse bias might exist. We weighted results to adjust for this possible bias, but weighting adjustments are imperfect.
Third, we cannot verify the accuracy of our respondents’ attitudes or behavior. We did not ask why physicians engaged in behavior that runs counter to professional norms and principles. In certain circumstances, physicians might have believed that their behavior was fully justified.
Fourth, although we found intriguing associations between communication concerns and physician gender, these results require cautious interpretation. Gender is closely linked to physician specialty, which we also included in our analyses. However, we did not have sufficient numbers of women in some specialties to include variables that explored the link between gender and medical specialty in our models.
Finally, to avoid burdening respondents and to achieve acceptable response rates, we restricted the number of questions about each professionalism domain in the multiple survey modules. We therefore did not ask respondents to provide details about their communication experiences. We also did not offer respondents opportunities to explain their responses or provide more-nuanced responses.
Exhibit 1 shows the demographic, professional, and practice characteristics of survey respondents. Most were male and white or Asian, had graduated from a US or Canadian medical school, and had practiced for more than twenty years.
|Characteristic||Number of respondents||Weighted percent of respondents|
|Race or ethnicity|
|White or Asian||1,648||89.6|
|Location of medical school|
|United States or Canada||1,331||72.2|
|Years in practice|
|Fewer than 10||210||12.4|
|30 or more||579||29.4|
|University or medical school||117||5.5|
|Hospital or clinic||343||18.8|
|Solo or two-person practice||401||22.0|
SOURCE Authors’ survey of US physicians in seven specialties.
Communication Attitudes And Behavior
Exhibit 2 shows respondents’ communication attitudes and behavior. The vast majority of the respondents completely agreed that physicians should fully inform patients about the risks and benefits of interventions, never tell a patient something that is not true, and never disclose confidential information to unauthorized persons. However, more than one-third of respondents did not completely agree that it was necessary to disclose all serious medical errors to affected patients or that it was important to disclose to patients any financial relationships with drug and device companies.
When asked about their own behavior in the previous year, more than a tenth of respondents said that they had told an adult patient or child’s guardian something that was not true. Almost one-fifth said that they had not fully disclosed mistakes to patients for fear of being sued. More than one-quarter reported revealing unauthorized health information about a patient. And more than half said that they had described a prognosis more positively than the facts warranted.
Associations With Physician And Practice Characteristics
Appendix Exhibits 3 and 436 display results from multivariable logistic regression models predicting responses that are concordant with the charter for the five communication attitudes and four behavior questions, respectively. Except for the behavior of fully disclosing mistakes, all attitudes or behaviors were significantly associated with one or more physician or practice characteristics, although no overarching pattern appeared.
Physician gender was significantly associated with four of the nine responses: those concerning never telling patients something untrue, fully describing benefits and risks, disclosing financial relationships, and never having told an untruth in the prior year. In all four instances, women were more likely than men to provide responses consistent with charter principles.
Race or ethnicity was significantly associated with two responses—attitudes about never telling an untruth and never disclosing confidential information. In both instances, underrepresented minorities were more likely than white or Asian respondents to report attitudes consistent with charter commitments.
Graduation from medical schools outside the United States or Canada was significantly associated with four responses. In three cases (never telling untruths, never disclosing confidential patient information, and never having revealed confidential information), graduates from these medical schools were more likely than graduates from US or Canadian schools to respond in ways that were consistent with charter precepts. In one case (attitude about disclosing financial relationships with drug and device companies), the difference between the two groups was inconsistent.
Physician specialty was significantly associated with four of the nine responses, although without consistent patterns relating to charter obligations by different specialties. For instance, general surgeons and pediatricians were most likely to completely agree about needing to disclose all serious medical errors to patients, while cardiologists and psychiatrists were least likely to report this attitude (). However, responses concerning self-reported actual behavior in disclosing errors found no significant differences by specialty.
In contrast, anesthesiologists, general surgeons, and pediatricians were most likely to report never having described patients’ prognoses in more positive terms than warranted, while internists and psychiatrists were least likely (). Cardiologists and general surgeons were most likely to report never having told patients an untruth in the previous year, while pediatricians and psychiatrists were least likely to report never having told untruths ().
Practice characteristics, length of time in practice, and local malpractice environment were sporadically associated to a significant degree with reported attitudes and behavior. For example, more physicians (78.1 percent) practicing in universities or medical centers completely agreed with the need to report all serious medical errors than physicians in solo or two-person practices (60.5 percent; ).
Among physicians practicing in regions with the lowest third of malpractice claim rates, 68.9 percent completely agreed that physicians should fully disclose financial relationships with drug and device manufacturers to patients, compared with 60.9 percent of physicians in regions with the highest third of malpractice claims ().
Despite widespread acceptance of communication principles and commitments by professional organizations, substantial percentages of US physicians did not completely endorse these precepts, and many reported behaving in ways that deviated from to these norms. Overall, approximately one-third of physicians did not completely agree with the need to disclose serious medical errors to patients, almost one-fifth did not completely agree that physicians should never tell a patient something untrue, and nearly two-fifths of physicians did not completely agree that they should disclose their financial relationships with drug and device companies to patients.
Perhaps more important, one-fifth of physicians reported not fully disclosing medical mistakes to patients because of fears of lawsuits, and just over one-tenth said that they had told patients something untrue in the previous year.
The survey results suggest that many physicians do not completely support the charter requirements related to communication with patients. An alternative interpretation is that treating support for the charter precepts as “black or white”—physicians either do or do not completely endorse and adhere to these principles—fails to recognize complexities of patient-physician communication in everyday practice.
Furthermore, although this survey was anonymous, many respondents were probably aware of which responses best fit expected professional principles and obligations. These findings are therefore likely to underestimate how often physicians hold these attitudes or how often this behavior actually occurred. All of these interpretations raise questions about how often and in what contexts patients are receiving complete and accurate information from their physicians that can be essential to patient-centered care.
Despite the relative clarity and unambiguous language of the charter precepts, many factors can affect how and what physicians communicate to patients. Some might argue that knowing when to breach or bend these rules—when individual patients require a different approach—constitutes clinical wisdom and true patient-centeredness. For instance, providing a patient with every detail about his or her case is rarely feasible, nor is it necessarily desirable. Physicians must sort through often contradictory and confusing information as their clinical assessments evolve and finally crystallize. Conveying many details—some of which may be erroneous—to patients might not prove helpful.
Some physicians may wonder about revealing errors to certain patients if no serious harm resulted from them. Nonetheless, informing patients fully about medical errors can reduce anger and lessen patients’ interest in bringing malpractice lawsuits.17 Plaintiff depositions from forty-five lawsuits found that 71 percent cited “relationship issues,” including doctors’ devaluing patients’ views and not understanding their perspectives.37
Findings about telling the truth probably reflect complex interpersonal and cultural influences.40 Studies of communication with gravely ill patients show that patients prefer honest and accurate information, delivered with empathy and understanding by clinicians, even when prognoses are dire.41 Nonetheless, some physicians might not tell patients the full truth, to avoid upsetting them or causing them to lose hope.22,42⇓–44 Especially in the context of life-threatening illness, physicians might not tell patients the complete truth because of lack of training, time limitations, uncertainty about prognostic accuracy, family requests, and feelings of inadequacy about their medical interventions.44
Physicians’ sex and race or ethnicity were significantly associated with fewer than half of the communication areas we assessed (four of the nine questions for sex and two for race or ethnicity). However, in all six instances, the direction of the differences suggested that women were more likely than men, and underrepresented minorities were more likely than whites or Asians, to follow charter principles.
As underrepresented groups within the medical profession, which white males have historically dominated, women and minority physicians may feel more compelled to comply with articulated professional precepts. They may seek to avoid attitudes or behavior that put their professional standing at risk. For example, physicians from underrepresented minority groups were significantly more likely than whites or Asians to indicate that physicians should never disclose confidential information to an unauthorized person.
Differences in communication attitudes and behavior by specialty were not surprising, although no consistent patterns appeared. The invasive nature of surgical interventions shapes relationships between patients and general surgeons: Patients must have high levels of trust in their surgeons to be comfortable moving forward with surgery.45 Therefore, it is reassuring that general surgeons were significantly less likely than physicians in most other specialties to report having told patients an untruth, and surgeons were significantly more likely than other physicians to agree with the need to disclose all medical errors to patients.
We did not ask about specific circumstances in which surgeons gave patients untrue information. It is possible that surgeons did so for reasons that they viewed as justifiable. For example, surgeons could have given patients untrue information to prevent patients from worrying. Whether this rationale excuses the provision of erroneous information would be likely to generate debate and requires additional study.
Clearly, further research that gathers more in-depth information from physicians is required, so that we can better understand the reasons underlying their failure to fully support professional communication principles. Examining circumstances in which physicians feel justified in not complying fully with charter precepts is especially important.
Another problematic area involved revealing information about financial relationships with drug and device companies. Substantial numbers of physicians did not fully support disclosing these potential conflicts of interest to patients. This finding is important, given the enactment of the Physician Payment Sunshine Act of 2009, which requires companies to begin reporting payments to physicians in excess of $10 by March 2013.
Once these data become public, some physicians will probably encounter patients who wish to discuss potential financial conflicts of interest. Physicians who do not support public disclosure might resist communicating this information to inquiring patients or might make these conversations difficult. Monitoring the effects of the law on this aspect of physician-patient communication will be important.
Despite the limitations of our study described above, our findings raise concerns that some patients might not be receiving complete and accurate information from their physicians. The effects of these communication lapses are unclear, but they could include patients’ lack of information needed to make fully informed decisions about their health care.
These information deficits could contribute to poor-quality care by making patients less able to make health care decisions that reflect their values and goals. Our findings raise questions about whether patient-centered care is broadly possible without more widespread physician endorsement of the core communication principles of openness and honesty with patients.
External financial support was provided by the Institute on Medicine as a Profession at Columbia University.
ABOUT THE AUTHORS: LISA I. IEZZONI, SOWMYA R. RAO, CATHERINE M. DESROCHES, CHRISTINE VOGELI & ERIC G. CAMPBELL
In this month’s Health Affairs, Lisa Iezzoni and coauthors present data from a 2009 survey of practicing physicians nationwide to assess how broadly doctors agree with and follow standards requiring openness and honesty in communication with patients. For example, the vast majority of respondents agreed that physicians should fully inform patients about the risks and benefits of interventions. However, nearly two-fifths did not completely agree that they should disclose their financial relationships with drug and device companies to patients. The authors say that their findings raise doubts about the viability of patient-centered care without more widespread physician endorsement of openness and honesty with patients.
Iezzoni is a professor of medicine at Harvard Medical School and director of the Mongan Institute for Health Policy at Massachusetts General Hospital. In addition to her academic and medical affiliations, Iezzoni is a member of the Institute of Medicine and serves on the editorial boards of Health Services Research and theDisability and Health Journal. She received the Distinguished Investigator Award from AcademyHealth in 2010. Iezzoni earned her medical degree from Harvard Medical School.
Sowmya Rao is an associate professor in the Quantitative Health Sciences Department at the University of Massachusetts Medical School, in Worcester, and a senior statistician at the Center for Health Quality, Outcomes, and Economics Research at the Bedford Veterans Affairs Medical Center, in Bedford, Massachusetts. Her main research interest is the development of statistical methods that use survey data to address questions of epidemiologic and public health importance.
Rao received her doctorate in biostatistics from Boston University and completed a postdoctoral fellowship at the National Cancer Institute.
Catherine DesRoches is a senior researcher at Mathematica Policy Research, with fifteen years of health services and survey research experience. Prior to joining Mathematica in 2010, she spent ten years as a researcher and faculty member at Harvard Medical School and the Harvard School of Public Health. Her work with Iezzoni, Campbell, Rao, and Vogeli was awarded the Professionalism Article Prize in 2010 by the American Board of Internal Medicine Foundation.
DesRoches received a doctorate from Columbia University’s Mailman School of Public Health.
Christine Vogeli is a health services researcher at the Mongan Institute. Her research focuses on physician professionalism and health care quality. She received a doctorate in health policy and management from the Johns Hopkins Bloomberg School of Public Health.
Eric Campbell is an associate professor of medicine at Harvard Medical School and director of research at the Mongan Institute. Over the past decade, his research has focused broadly on professionalism in medicine and has empirically examined issues related to conflicts of interest, self-regulation, care for the poor, and participation in civic activities.
Campbell received a master’s degree in education, with a specialization in adult education, and a doctorate in higher education policy from the University of Minnesota.
- ↵To access the Appendix, click on the Appendix link in the box to the right of the article online.