Sankar P, Jones NL. To Tell or Not to TellPrimary Care Patients’ Disclosure Deliberations. Arch Intern Med. 2005;165(20):2378-2383. doi:10.1001/archinte.165.20.2378
Copyright 2005 American Medical Association. All Rights Reserved. Applicable FARS/DFARS Restrictions Apply to Government Use.2005
Literature on patient disclosure has highlighted policy proposals, practitioner attitude and practices, and issues specific to already identified vulnerable populations (eg, human immunodeficiency virus–positive patients and adolescents). We investigated the topics and concerns that led less-studied primary care patients to deliberate disclosure to their physician and the range of actions taken following such deliberation.
This descriptive qualitative study used semistructured in-depth interviews with 85 self-selected female primary care patients recruited from 2 hospital-associated primary care practices and through advertisements in local newspapers.
Topics that led subjects to deliberate over disclosure included sex, sexually transmitted diseases, reproductive health, drug use, mental health, and domestic circumstances. Concerns prompting this deliberation fell into 2 categories. The first, extraindividual concerns, included fear of information circulating to employers or insurance companies and the need to disclose for adequate treatment. The second, intrapersonal concerns, centered on feelings of comfort, embarrassment, or shame in the act of disclosing during the clinical encounter. While the majority of women did eventually disclose all or part of the information in question, a significant minority did not disclose. There were no significant demographic differences associated with patterns of disclosure.
Primary care patients have concerns that affect and sometimes limit disclosure. Physicians are encouraged to explore patients’ concerns about the uses and circulation of medical information, as well as anxiety over anticipated shame or embarrassment at the moment of information disclosure to physicians. This is especially important in primary care settings, where failure to disclose can mean lack of referral to needed specialized care.
Effective treatment relies on patient disclosure of relevant information. Medical confidentiality protections are meant to encourage disclosure, and typically patients do disclose what is needed. However, under what circumstances patients decide not to disclose and whether these decisions relate to confidentiality protections, or to other factors, is not well understood, especially for the primary care patient.
The literature on patient self-disclosure is small, but varied. One type offers policy proposals or reflections on how best to protect certain classes of patients, such as those undergoing genetic testing,1,2 those dealing with end-of-life issues,3 and adolescents.4 Another reports on experiments with innovative techniques to facilitate patient disclosure on socially sensitive topics, such as sexual activity among adolescents,5,6 domestic violence,7 and incest.8 Related to this research, another set of articles9- 11 reports on studies that examine the practices and beliefs of physicians and nurses that might influence patient disclosure, an especially important factor in certain circumstances, such as emergency department disclosure of domestic abuse. Research on the attitudes of patients toward disclosure has focused on the reticence of people enrolled in treatment programs, such as for human immunodeficiency virus or drug rehabilitation, to disclose continuing information about their sexual or drug-related conduct.12- 14
Research on primary care has explored which topics are most likely to trouble patients, and has found that disclosure concerns cluster around sexual problems15- 19 and mental health.20- 23 Most of these studies sought to uncover barriers to disclosure, finding most commonly that nondisclosure was associated with a belief that the problem was not worthy of discussing with a physician,15,17,23 with certain characteristics of the physician, such as age or sex, or with the patient’s sense that the physician is not interested.17,20,22 Conversely, studies of patient characteristics that are associated with treatment seeking and disclosure found that lack of embarrassment in front of physicians18 and a positive attitude regarding disclosure in general21 favors disclosure.
Less attention in studies of primary care patients has been paid to the range of concerns and the types of medical information that motivate disclosure deliberations. Also lacking is a description of what patients do following deliberation. While some might see their choice as only telling or not telling, others may develop alternative strategies. Understanding the patient’s perspective on information disclosure is a key ingredient in encouraging patient openness and collaboration in health management and is particularly important for the primary care physician, who acts as a gatekeeper by referring patients for more specialized physical or mental health care.
To contribute to practitioners’ understanding of patient disclosure concerns, we conducted in-depth open-ended interviews with 85 women recruited through primary care clinics and local newspaper advertisements. The interviews explored models, definitions, and concerns about medical confidentiality and how those concerns affected disclosure behavior.
Data for this study were collected during in-depth open-ended interviews, lasting between 35 and 90 minutes, with 85 women aged 18 to 75 years (average age, 35 years). Women were recruited through 2 primary care clinics associated with large hospitals and through advertisements in 3 local newspapers in southeastern Pennsylvania and central New Jersey. Women 18 years and older were eligible to participate, and each was paid $25.
The study focused on women because of their frequent role as coordinators of family care, and because women on average have more interactions with physicians in clinical settings.24- 26 In addition, because of the relative lack of research about the patient’s perspective on these issues, limiting the research to women helped focus the inquiry and provide direction for future study. This choice carries the obvious disadvantage of being of limited relevance to men, especially important because the scant research available suggests that men and women might treat sensitive health issues quite differently from one another.27,28 Nevertheless, the data herein provide a foundation on which to base subsequent broader research.
Interviews were conducted using an interview guide (available on request from the authors) that had been developed and extensively piloted in several health care settings. The validity of a questionnaire such as this one, which not only inquires into potentially sensitive topics but also into deliberations over whether to conceal or reveal such topics, is, by definition, difficult to establish. There is no observer-independent gold standard, because even the medical record could not stand as evidence that a person had or had not experienced some event if the person had chosen to conceal it. However, the circumstances of the questionnaire’s design partially addressed validation concerns. The questionnaire was initially formulated and tested in a clinic where we had already conducted 6 months of ethnographic research. That earlier investigation provided a preliminary sense of the kinds of issues encountered by this population and the range of conduct associated with disclosure. This provided an independent knowledge base that could be relied on to revise or eliminate questions that seemed to elicit inaccurate answers.29 Furthermore, the likelihood of respondents intentionally misreporting experiences or misrepresenting beliefs is considered lower in interviews that seek to document the concerns of subjects as patients in contrast to those that monitor individual conduct.30
The final interview guide included closed- and open-ended questions about women’s perspectives on medical confidentiality and their concerns over its possible breaches, and the role of these concerns in communication with physicians. Interviews were conducted in private rooms at the clinic or at a research or community center, by a person with a recent doctorate degree in anthropology and secondarily by one of us (P.S.). The institutional review board of the University of Pennsylvania and of the hospitals where the clinics were located approved this research.
All interviews were transcribed and analyzed using computer software (QSR NUD*IST v.4).31 The research team used multilevel consensus coding, requiring an iteratively developed coding scheme and multiple reviews by paired coding teams. Details of this method have been published elsewhere.32
Transcript coding revealed a set of common stories among subjects in which they recount deliberations over whether to tell their physician or a family member some piece of information perceived as sensitive. Coding identified several hundred such anecdotes, which we labeled tell/not tell (TNT) stories. Limiting the analysis to those TNT stories concerning physicians eliminates stories from 3 participants because their deliberations concerned disclosure to family or friends only. Furthermore, although nearly all women told at least 2 TNT stories, some told more. To create a sample that did not substantially overrepresent the latter group, analysis herein is limited to a maximum of 3 TNT stories per person. Stories were entered into the database after being assigned random numbers. For women who shared more than 3 TNT stories, we included the first 3 listed for them in the database. Our final data set, therefore, is 187 TNT stories from 82 women. For each TNT story, we assessed the topic and the reason for deliberation, and analyzed the action taken after deliberation.
Table 1 provides the age, education, and income ranges of the 82 women providing TNT stories. The income distribution and the racial or ethnic identities of the participating women are comparable to that of the regional population, although our sample is slightly younger and more educated.33 (All 3 of the women eliminated from analysis because they had TNT stories exclusive to family or friend disclosure were white and had annual incomes between $41 000 and $75 000; 2 had attended some college, and 1 was a college graduate; 1 was between 23 and 45 years old, and 2 were older than 46 years.)
Women listed a range of topics that prompted them to deliberate over disclosure. The most common, in order of frequency named, are as follows: sex and sexuality, sexually transmitted diseases, reproductive health, drugs, mental health, and domestic issues. Additional topics mentioned less frequently included incontinence, memory loss, diet and personal appearance, genetics, cancer, diabetes mellitus, and minor medical problems. Many stories did not address a specific topic, but instead described customary disclosure practices and concerns, a category labeled as “general.” Table 2 provides the full list of topics triggering deliberation, in order of frequency mentioned. The remainder of the results report only data associated with topics named by 20% or more of the women.
Disclosure deliberations were triggered by concerns that fell into 2 broad categories: (1) future and extraindividual and (2) immediate and interpersonal. These categories emphasize the timing and the locus of concern evoked for the patient by information disclosure, whether at the moment of disclosure or later and whether internal to or outside of the consultation. Concerns coded as future and extraindividual relate to where and to what other parties information might travel after the appointment ends. These concerns typically related to continuity of care, insurance eligibility, employment records, and, for younger women, parental access to information. Within this category, the following subcodes were identified: information flow, diagnosis/treatment, and “none of your business.” Definitions of subcodes and sample passages are presented in Table 3.
Concerns coded as immediate and interpersonal focused on the emotional experience of self-revelation in the face-to-face physician-patient encounter. These concerns highlight the anxiety triggered by having to determine whether to reveal to the physician information perceived as intimate or embarrassing. Immediate and interpersonal TNT stories anticipated feelings of shame or being judged and expressed the need to feel at ease before opening up. Within this category, analysis revealed 2 subcategories: comfort/discomfort and shame, embarrassment, or fear of being judged (Table 3).
Approximately 60% of TNT stories were motivated by future and extraindividual considerations. Of these stories, approximately equal numbers were coded to information flow and to diagnosis/treatment, followed by fewer to none of your business. In the TNT stories motivated by immediate and interpersonal concerns, more were coded to shame than to comfort.
When subjects weighed the choice to disclose information, they described several possible actions, ranging from proceeding directly with full disclosure, to telling part of their information, to withholding it entirely or lying. We divided the described actions into 3 categories: exchange, qualified exchange, and not exchanged. Exchange encompasses full disclosure, and not exchanged includes not telling and lying. Qualified exchange is the most heterogeneous category, and includes telling the story using a friend as the subject, delaying disclosure, telling partial information, and seeking care anonymously. While most TNT stories concluded with full or qualified exchange, 40% ended in nondisclosure.
To provide a more comprehensive account of study results, Table 4 provides examples of TNT stories, highlighting the relationship between concern prompting deliberation and the action taken.
Most of the TNT stories did not include specific statements about how physician conduct or characteristics directly influenced patient willingness to disclose. However, those that did (30 of 187) fell into 2 distinct groups. The first set of comments refers to features over which physicians have no control; however, awareness about their influence on patient willingness to disclose might alert physicians to the increased possibility of nondisclosure. These features include certain demographic similarities between the patient and the physician, such as age, language, sex, and race or ethnicity, and length of time or number of times a patient has seen a physician. These comments were not offered as necessary conditions for disclosure, rather only that women reported feeling “very open” with physicians with whom they shared these sorts of characteristics, and feeling more comfortable increased the likelihood they would disclose information.
The second set of comments could be categorized as concerning the physician’s communication style, specifically that disclosure relied on the patient feeling confident that the physician was actively listening. Details in these comments included the following: inviting the patient to speak, not rushing the patient, and gently encouraging the patient if the physician feels something is not being disclosed. One woman explained how physicians might encourage patients to talk: “Like my internist . . . if the doctor is intelligent and sensitive, they will try to get those issues without sounding like it’s a big deal. Like it’s something natural that people go through. So, communication is very important with a doctor, and sometimes it’s not easy to have that because the doctors are rushing you and you have to wait so long.”
Patients should and do exercise some control over what medical information they share with physicians. Recognizing that time is limited and that not everything can be reviewed, patients can reasonably choose to abbreviate. This study suggests that several topics and concerns limit patient willingness to disclose, draws attention to concerns that extend beyond those usually addressed by standard confidentiality protections, and indicates possible physician responses that may increase disclosure rates.
This study examined statements by women about how they weighed the decision to disclose medical information to physicians. It found several topics, including reproductive health, domestic circumstances, mental health, sexually transmitted diseases, and substance use, about which women deliberated disclosure. These topics are recognized as sensitive and their disclosure has been studied, although typically within populations already characterized as vulnerable, such as adolescents,34 people with mental illness,35 and those considering genetic testing.36 This study examined disclosure concerns in a primary care setting, where refusal to disclose not only impedes routine health care, including preventative health, but hinders recognition of the need for specialized treatment.23,37,38 This is particularly true for the increasing number of patients insured through managed care programs, in which referrals to specialists are triggered by disclosure to the primary care provider.
There are potential limitations to these findings. First, we studied a small self-selected sample of primary care patients who may not be representative of the general population, and we included only women. Second, we relied on self-reported past behavior. Third, questions that ask people to disclose to an interviewer what they have withheld from a physician are difficult to validate. Still, the findings demonstrate important concerns that previously have been overlooked or underestimated.
Four findings deserve particular attention. First, of the TNT stories examined, most resulted in full or partial disclosure. While this suggests optimism that patients are communicating relevant information to physicians, the many qualified exchanges indicate caution. Qualified exchange is the most heterogeneous category, and includes some actions that, although ultimately resulted in exchange, did so either at the cost of delay or by going to another physician. While the latter can be an appropriate solution, it can also result in multiple fragmented medical records, which impedes effective care.39
Second, the many women who chose not to disclose also deserve attention. Women in our study did not disclose a range of topics and behaviors, including sex, sexually transmitted diseases, mental health, and domestic issues. Lying about or not disclosing smoking or the full extent of smoking was one of the most common nondisclosure TNT stories. This finding, in combination with research showing that physicians overestimate their ability to detect smokers among their patients,40 suggests that smoking disclosure merits extra attention.
Third, the none-of-your-business category of disclosure, although small in number, is notable for the certitude with which the subjects delivered the judgment that certain questions sought information that was unrelated to their medical care. They were confident that sexual history, history of sexually transmitted diseases, and alcohol and tobacco use were not germane to their current ailment and, thus, not only did they not reveal the information, in many cases came to think of their physician as intrusive for asking.
Finally, the many women citing immediate and interpersonal factors, such as feelings of shame, as limiting disclosure suggests that disclosure concerns extend beyond the legalistic or bureaucratic framework represented by the Health Insurance Portability and Accountability Act,41 which focuses on information circulation beyond the clinical setting.
These findings have practical implications. Most important, they confirm the need to communicate clearly and repeatedly to patients the distinctiveness of the medical setting. In this setting, topics are discussed that are socially proscribed elsewhere. The degree of social disapproval associated with some topics, typical to the medical setting, suggests that the preference for candor might need to be explicitly addressed. Patient comments that physicians might try gently encouraging disclosure support this. Effort should be made to recognize and mitigate potential embarrassment or shame, and the need to do so impressed on clinic staff as well. Furthermore, while the reason for collecting certain medical facts might seem clear to the physician, this research suggests that some patients, unaware of these justifications, withhold requested information. To counter this behavior, physicians might include a more complete explanation of the reason motivating a question, especially concerning topics emphasized in these findings, such as sexual history.
We are not suggesting that more information or unfettered disclosure is necessarily better than selective reporting. Patients, as physicians, need to assess what is essential to communicate in the time provided and what can be set aside. The questions put to subjects and the data reported on herein, however, started from a more narrow premise. In the case of a patient who was aware of something she thought medically relevant or given a physician specifically inquiring about some piece of information, what factors did the patient consider before going ahead and sharing that information?
Also, emphasizing the importance of physician conduct to facilitate patient disclosure, as we have, is not meant to suggest that physicians bear sole responsibility for factors that hinder patient openness. Social trends that have eroded traditional confidentiality practices and social beliefs that characterize certain health-related topics as offensive or inappropriate for discussion contribute significantly to patient concerns. But, it is also clear that physician effort to assess patient concerns about disclosure in the medical setting and to allay concerns when feasible could benefit patient treatment and possibly strengthen physician-patient trust.
Correspondence: Pamela Sankar, PhD, Center for Bioethics, University of Pennsylvania, 3401 Market St, Suite 320, Philadelphia, PA 19104 (firstname.lastname@example.org).
Accepted for Publication: June 22, 2005.
Financial Disclosure: None.
Funding/Support: This study was supported by a grant from the Rockefeller Brothers Fund (formerly the Charles E. Culpeper Foundation), New York, NY.
Role of the Sponsor: The funding body had no role in data extraction and analyses, in the writing of the manuscript, or in the decision to submit the manuscript for publication.