Confidentiality and Truth Telling

In 1982, a study by Siegler, an internal medicine physician, found that roughly 75 health care personnel, including 6 attending physicians, 20 nurses, 15 students, 12 residents, 4 financial officers, 4 hospital reviewers, and others, legitimately had access to the chart of one patient during the course of a single, brief hospital stay (Siegler 1982). The exasperated patient remarked, “Perhaps you should tell me just what you people mean by ‘confidentiality’!” In this day of managed systems of clinical care, with computerized medical charts, laboratory reports, and billing procedures and broad-scale sharing of data for research, it would be nearly impossible to develop an accurate estimate of the vast number of individuals with potential access to an individual patient’s clinical records.

Empirical work in the area of confidentiality is very limited, but early findings suggest that efforts to protect patient confidentiality are variable and difficult to achieve. In a self-report, structured interview study with 747 adolescents, 76% indicated that their confidentiality had been violated by a clinician in the past (Cogswell 1985). Female participants in this study identified confidentiality breaches far more often than did their male peers, especially in regard to reproductive health issues (72%) as opposed to general health issues (28%). In a novel study examining the actual behaviors of health professionals, inappropriate disclosures of patient information were noted on 14% of 300 observed hospital elevator rides at one institution (Ubel et al. 1995). A questionnaire study with 177 patients, 109 house staff, and 53 medical students further revealed that patients expect greater levels of respect for patient confidentiality than actually are present in the training hospital setting (Weiss 1982). For example, whereas only 17% of patients thought that it was common practice for medical personnel to talk with their spouses or partners about patient cases, 51% of house officers and 70% of students reported this perception. Only 9% of patients thought that clinicians would talk about patients as “interesting stories” at parties; 45% of students and 36% of house officers, however, felt that this was common.

Significant problems related to confidentiality have been shown to exist with respect to mental health treatment specifically. Complaints regarding confidentiality are a common trigger for investigations and disciplinary actions by the ethics committees of the American Psychiatric Association and the American Psychological Association (Pope and Vetter 1992). Nine percent of psychiatric outpatients and 15% of psychological outpatients in two other studies indicated that their therapists had breached confidentiality in the past (Wettstein 1994). A series of studies on group therapists’ attitudes and behaviors found that whereas three-quarters of group therapists affirmed the importance of confidentiality, only 32% instructed their group participants about appropriate safeguards for patient disclosures (Roback et al. 1996). More than half (54%) of 100 members of the American Group Psychotherapy Association randomly surveyed reported that confidentiality had been breached by group members at least once in their practices.

Studies of patients with mental illness have also shown that about half experience significant concern about confidentiality but only a small proportion are aware of confidentiality measures that exist for their protection, e.g., surrounding documentation of sexual health issues or substance abuse treatment (Lindenthal and Thomas 1982; Wettstein 1994). Such findings are especially ominous in view of the fact that fears of confidentiality breaches have been repeatedly shown to prevent patients from seeking or engaging fully in necessary mental health care (Roback et al. 1996; Roberts et al. 1999; Wettstein 1994). A survey of 76 psychiatric patients and 76 community members revealed, for example, that one-quarter to one-half felt that the risk of a psychiatrist divulging personal information had prevented them “to a great extent” or “to some extent” from seeking mental health treatment.

In general medical care, the clinical value of protecting sensitive patient information has also been demonstrated empirically. In a study of 56 adult women patients, 91% indicated that “security, trust, and confidentiality” were among the top five qualities desired in a physician when seeking out advice on sexual health matters (Metz and Seifert 1988). Participants in a survey of 102 self-identified gay, lesbian, and bisexual youth ages 18–23 years reported far greater willingness to discuss their sexual orientation, sexual health concerns, and sexually risky behaviors after receiving accurate information and reassurance about confidentiality safeguards related to their care (Allen et al. 1998). An anonymous survey of 1,295 high school students revealed that 25% would forgo necessary treatment if their parents “might find out” (Cheng et al. 1993). Such work underscores the importance of confidentiality protections to the use of health services and the effectiveness of clinical assessment and treatment planning.

Relatively few empirical studies exist in the area of truth telling in clinical care. Accurate disclosure of medical mistakes is an area that has begun to receive tremendous attention, however (Rubin 2000; Institute of Medicine 1999). Until relatively recently, only a few case reports had entered the literature (Baylis 1973; Finkelstein 1974). In 1997, Sweet and Bernat surveyed 150 randomly selected medical students, house officers, and attending physicians regarding two vignettes in which an “error of commission” occurred (i.e., a physician medication error resulting in coma, seizures, and enduring pain and a physician medication error resulting in patient death) and one vignette in which an “error of omission” occurred (i.e., a referral physician’s failure to diagnose a patient’s cancer, leading to unnecessary paralysis). A majority of participants (95% and 79%) indicated that they would tell the patient or his family truthfully about the mistake in the first two scenarios, but only 19% indicated that they would report the error of the colleague in the third scenario. Given that preventable adverse medication errors occur in at least 2% of hospital admissions, the findings of this study are very relevant and important for clinicians (Sweet and Bernat 1997).

In terms of truthfully delivering “bad news” such as a poor prognosis diagnosis to patients, a structured interview study of 32 U.S. physicians related to care of the terminally ill revealed that only 47% would tell the patient his diagnosis explicitly, whereas 22% would “use a euphemism” and 31% would not tell the patient at all (Miyaji 1993). In a recent cross-cultural study, 200 elderly Los Angeles residents were surveyed regarding truth-telling issues in diagnosing and treating terminal illness. Korean American (47%) and Mexican American (65%) participants were much less likely than Anglo (87%) and African American (88%) participants to believe that a patient should be told the diagnosis of metastatic cancer, and strong cultural differences were also found in opinions regarding the role of families and the relative influence of clinicians in making end-of-life decisions (Blackhall et al. 2001). A study of 677 medical geneticists in 18 nations similarly revealed considerable variability in their responses regarding whether they would 1) disclose the diagnosis of XY genotype in a female gender patient, 2) indicate which parent carries a translocation causing Down syndrome, or 3) reveal a diagnosis of Huntington’s disease or hemophilia A against the wishes of the patient who was tested (Wertz et al. 1990). Interestingly, ethical withholding information from patients was the top-ranked topic (75% of respondents) deserving more curricular attention in a study of 181 psychiatry residents at 10 training programs (Roberts et al. 1996).

Significant disagreement exists regarding standards for ethical disclosure of patient information even in situations in which reporting is mandatory. Indeed, one recent study showed that up to 75% of clinicians would elect not to report a case of suspected child abuse, despite the fact that this “omission” would be in violation of the law (Wettstein 1994). In terms of falsely presenting or withholding information, a U.S. survey with 211 physician participants published in 1989 revealed that a majority were willing to misrepresent a screening test as a diagnostic test in order to obtain insurance reimbursement. One-third said that they would offer incomplete or misleading information to family members regarding a medical mistake resulting in the death of a patient. Only 25% indicated that they “never” employed deception of any form in patient care, and 27% agreed with the statement that patients “expected” them to “utilize deception” for patient benefit (Novack et al. 1989). A mail survey of 510 Kansas family and general practitioners revealed that small community size was a critical problem affecting patient privacy and physician behaviors related to confidentiality and truth telling. These physicians’ techniques to protect patient confidentiality included: speaking with office staff about confidentiality (88%), omitting details from the medical record (69%), charting the importance of confidentiality (46%), omitting details from insurance forms (37%), failing to notify public health officials (21%), and misrepresenting facts in the medical record (6%) or on insurance forms (5%) (Ullom-Minnich and Kallail 1993).

Finally, only limited data have been reported on the ethical imperative to tell the truth as it relates to “whistle blowing.” For instance, case data published by the Office of Research Integrity within the U.S. Public Health Service revealed 986 allegations of research misconduct in the 5-year period 1993–1997 inclusively (Price 1998). In 1996, 3,653 disciplinary actions against physicians occurred due to alleged violations of law, ethics, or practice standards in the United States. Beyond such anecdotal reports, however, no data currently exist on the failure to report unethical colleague behavior in either the research or the clinical domain.

There are circumstances when the risks associated with nondisclosure of patient information in mental health care are very great and outweigh the individual’s privilege of confidentiality. Reporting clinical findings becomes ethically and legally justified in situations where there is a serious and imminent threat of physical harm to an identifiable and specific person, where breaking the confidence is likely to do good and to prevent harm (e.g., protection of the intended victim), and where other efforts to address the situation have failed or are insufficient clinically or legally. This is the basis for mandatory reporting in relation to the diagnosis of communicable diseases, discovery of gunshot or other crime-related wounds, suspicion of child or dependent elder abuse, and drivers who are dangerous due to serious mental illness, epilepsy, or other conditions.

The landmark legal case of Tarasoff v. Regents of the University of California, 1974 and 1976 serves as a dramatic illustration. In this tragic case, a young woman named Tatiana Tarasoff was murdered by a graduate student, Prosenjit Poddar, who stabbed and shot her and who, in the course of psychiatric care, had previously disclosed his intention to kill Ms. Tarasoff. Although Mr. Poddar’s threats were taken seriously by the clinicians involved in his treatment, efforts to commit him involuntarily for inpatient psychiatric care and to retain him in outpatient treatment by the University of California failed. Though his attorneys had argued that he had diminished capacity due to mental illness, the Superior Court of Alameda County, California, found Mr. Poddar guilty of second-degree murder in 1974. This verdict was overturned as the result of errors made by the judge in instructing the jury. Mr. Poddar was subsequently allowed to return to India. Because of the immense controversy it caused, the case was re-reviewed by the California court system in 1976, leading indelibly to the clinician’s legal “duty to protect” potential victims of threatened violence.

From an ethical perspective, the tension between the professional principles of autonomy (i.e., furthering the rights and beliefs of patients) and nonmaleficence and beneficence (i.e., ultimately protecting the well-being of individuals) was central to the Tarasoff case. With the Tarasoff legal decision, fulfilling the responsibility to protect innocent others and society at large in circumstances of life-endangering patients became the predominant and enduring ethical imperative. Indeed, forensic issues often are predominant in making confidentiality decisions in clinical care. Specifically, there are five situations in which there are significant legal precedents suggesting that it is justifiable to disclose a patient’s personal information without explicit permission:

Within these legal parameters, however, there are a number of ethical judgments that must be made. Consider, for instance, the case of a hospitalized delusional patient who expresses homicidal intent but without reference to a specific individual. The patient is under close supervision, has no violence history, and is accepting medication. Is there any immediate need to take action confining the liberties of the patient or reporting him to local officials? How about if the psychiatrist knows that the patient has a long-standing paranoid belief about his next-door neighbor and has made threats toward him in the past—is this sufficient to warrant action? What if the patient then leaves the ward without the knowledge or permission of the staff? Consider a second example related to the timing of a mandated report of suspected child abuse. How much data gathering should the clinician engage in before complying with legal imperatives to disclose information to state or county agencies? Which member of the multidisciplinary treatment team should make the report? How should the process around reporting the situation be documented in the parent’s and the child’s medical charts?

Confidentiality issues arising in more mundane, everyday situations are also complex. Consider the multidisciplinary team treatment planning situation: patient information disclosed or discovered in the context of a therapeutic relationship may be discussed openly by members of a large clinical care team. Patient information is also commonly entered into a written medical record or computer database. Similarly, in teaching settings, it is essential that trainees obtain guidance from their supervisors regarding all significant patient care matters. What confidentiality protections exist in these discussions? Is every aspect of the patient’s life history included? How should sensitive material (e.g., related to drug use, sexual behavior, violence history, or HIV status) be documented? What limitations are placed on access to the patient’s chart or to the computer database? What if the patient is also a part-time employee at the clinic where she is receiving care, because her group insurance policy requires this? Other routine situations in which confidentiality decisions arise include elective patient consultations; disclosures of patient data to insurance reviewers, managed care companies, or legal agencies; and appropriate disguise of “protected” patient information in clinical presentations and scientific papers. These widely ranging situations are often very difficult and may create significant ethical binds for clinicians.

The Health Insurance Portability and Accountability Act (HIPAA), originally introduced in 1996 and scheduled to transition fully into practice by April 2004, offers additional protections for patients’ personal medical records. According to the U.S. Department of Health and Human Services, the aim of this set of regulations was to increase confidentiality protections and to balance the privacy needs of individuals with societal needs, such as public health, medical research, and quality assurance efforts and greater accountability around health care fraud and abuse. Key features of HIPAA include limitations on the use of personal health information by agencies and institutions for reasons other than direct clinical care (e.g., restrictions on use of data for research purposes), increased access by patients to their personal health records, increased access to patient data by governmental entities under some circumstances, and strengthened rights of patients regarding explicit advanced consent procedures for the release of identifiable data.

Respect for individuals and their privacy, respect for the truth, and respect for the law are three general principles that guide the resolution of these kinds of clinical ethics dilemmas. Although the three principles may be in tension, as noted in the illustrations above, they often are congruent or can be combined harmoniously within the context of the therapeutic relationship of the patient. For example, when assessment of a family reveals evidence of emotional, physical, or sexual abuse of a child, mandatory reporting can be integrated constructively into the therapy by 1) informing the parents of the clinician’s responsibility to protect the child, including the legal duty to notify authorities about the situation; 2) clarifying what the possible consequences of report may be and addressing the parents’ fears; 3) inviting the parents to take responsibility for immediately reporting the situation in the presence of, and with support from, the clinician; 4) mobilizing additional clinical services to help the family cope in the crisis situation; and 5) offering reassurance that the family will not be abandoned by the caregivers. Through such efforts, it is possible to act therapeutically and to embody respect for individuals, the truth, and the law.

In general, strategies for resolving difficult dilemmas related to confidentiality and truth telling are severalfold. First, it is critical to inform patients of confidentiality issues early in the therapeutic relationship and to explain how the professional and legal obligations of the clinician fit within this relationship. It is important to offer accurate explanations of confidentiality protections and their realistic limitations, given the clinical care context. It is helpful to foreshadow for the patient how specific consent will be sought for “voluntary” disclosures of personal information—for example, in family meetings or to insurance company reviewers. This is especially crucial in discussing potentially stigmatizing health issues, such as mental illness symptoms, sexual behaviors, genetic information, HIV risks, and substance use patterns. In this same conversation, it is possible to introduce the topic of disclosure practices that arise due to legal imperatives, such as when threatening, neglectful, or harmful behaviors fall under mandatory reporting laws. When dealing with health data that might affect insurance or employment status, it is also valuable to clarify the patient’s understanding of the impact of disclosure and the clinician’s responsibilities for documenting and reporting data revealed in the course of clinical care. With some very difficult patients who may purposely seek to pressure or even intimidate patients and staff, for instance in some drug treatment programs or inpatient units, this material may be introduced very naturally into larger discussions of behavioral expectations and consequences within the treatment setting.

Second, gathering additional information and guidance from supervisors and colleagues is important when one is dealing with novel or especially difficult ethical dilemmas related to confidentiality and truth telling. For example, a clinician taking care of a patient with newly emerging substance abuse problems and a high-stress job involving the safety of others (e.g., air traffic controllers, military personnel) may wish to speak with a specialist in the area of chemical dependence to discuss indicators of more serious symptoms, appropriate treatment-compliance strategies, and professional obligations in the specific clinical context. In addition, to fulfill basic standards of care in this country, it is essential for all clinicians to provide patients with complete and timely information about their diagnoses, prognoses, and potential therapeutic options, including the option of no treatment at all. Any deviation from this “truth telling” ethics practice expectation requires careful justification, reflection, and documentation and should be undertaken only after consultation with knowledgeable colleagues.

Third, it is important for clinicians to think carefully about who has access to patient data and how such information is shared. Legally, patients “own” their personal information, whereas the caregiver’s institution owns the actual clinical chart or computerized record. As noted in Chapter 3, this position, which has been maintained historically is being eroded by the existence of large databases, some of which contain genetic data in addition to general health data, and by recent softening of regulations. Many individuals legitimately have access to the chart or computerized record in the clinical institutional setting (e.g., for direct patient care or quality assurance reasons), and when insurance releases have been obtained, the numbers of individuals with access to detailed information about individual patients can be very high. In answering the question of who should see or receive information from a patient’s medical record, the new HIPAA regulations will offer additional protections to patients when information is sought for reasons other than direct patient care.

Other ethically important issues arise when information is given to patients or legal guardians. In accordance with HIPAA regulations, the majority of states in this country have laws explicitly allowing patients and guardians to have direct access to the medical chart. Sharing of information with patients and families must be understood within a therapeutic framework, however. Under ideal circumstances, patients and family members should be observed and supported during the process of reviewing the record. Alternatively, many clinicians will choose to sit with the patient in order to summarize and interpret orally the often confusing, potentially disturbing entries in the chart. Because the clinic, hospital, or medical facility actually possesses the chart and is legally responsible for it, patients should not be allowed to remove the record from the immediate care site. For these reasons, whenever possible, a secure, supervised area should be provided for chart review in health care settings.

A fourth strategy in dealing with confidentiality and truth-telling dilemmas is adopting the practice of revealing only what is absolutely necessary in all interactions outside of the clinician–patient relationship. This is important in both voluntary and involuntary disclosures in which personal information is released to anyone other than the patient (e.g., family members, insurers). This is also important in gathering advice from clinical colleagues. In general, such information should not be shared with an employer, unless this is clearly arranged and formally consented to by the patient at the time of the evaluation or is subject to mandatory reporting laws (e.g., uncontrolled seizure disorder in a school bus driver, suicidality in an airplane pilot with major depression). With very few exceptions, the chart and the data contained therein should never be given over as a whole or “carte blanche” to anyone outside of the clinical care team or appropriate medical facility personnel. This is true even in cases in which there are legal imperatives (e.g., caring for victims of criminal assault such as rape), because the chart often contains a wide range of personal material that is not germane to the situation and that may be damaging to the patient if released. For these reasons, in order to comply with the law while also affording maximal protection to the patient, the clinician may find it necessary to provide only portions of the charted data or to review the relevant material verbally with the outside reviewer. Even though it may be very burdensome, the clinician’s ethical duty to protect patient information is not abdicated in such circumstances.

Fifth, it is important to look for ways to improve how patient information is dealt with in the clinical care setting and clinical system. Sensitizing health care staff to the complex issues surrounding patient privacy and offering constructive guidelines for the workplace are critically important interventions for ensuring patient confidentiality. Maintaining secure and appropriately detailed personal notes or perhaps “shadow” charts—with clear notations in the “main” clinical record, including reference to the existence of the parallel protected chart—is a reasonable solution that many clinicians may employ when caring for patients with stigmatizing illnesses. This practice may be prohibited in many settings. In all circumstances, the main clinical record must document all key information needed to care for the patient (Roberts et al. 2002). The development and implementation of appropriate procedures to protect patient information contained on computerized databases serves as a second example. Simple measures that may be adopted by health care facilities include use of passwords and encrypted codes that are periodically revised. Limiting access to those portions of the database that contain potentially stigmatizing information is also crucial. Encouraging employees to position their monitors so that the data presented on their computer screens are not visible to passersby may also be helpful on an individual-user basis. Similarly, efforts to minimize clinical interactions in public areas, such as hallways and elevators, will help prevent inadvertent breaches of patient confidentiality.

Finally, when dealing with threatening, abusive, or neglectful patient behavior that falls under mandatory reporting laws (e.g., harm to a child or a dependent adult) or with colleague incompetence or impairment, it is ideal to set up a safe and supportive situation in which individuals may report their own behavior directly. Individuals who are ill, whose judgment is compromised, and whose behavior is actually or potentially dangerous will benefit from a process in which they assume responsibility and take the initiative to help themselves.

There are a wide range of approaches to resolving subtle everyday patient care dilemmas that clinicians take related to confidentiality and truth telling (Table 1). Some approaches are good, and some are not so good, from the perspectives of the law and of ethical practice standards. Purposefully inaccurate documentation of patient information on an insurance form, for example, is not an acceptable ethical solution to the problem of safeguarding patient privacy. Nevertheless, clinicians may choose, ethically, not to document their speculations about a specific patient care situation if they are in the process of gathering information and expertise that may help clarify the issues involved. In some cases, however, even this may be ethically problematic—for example, when legal imperatives relate to clinicians’ best judgments in the absence of complete information, such as mandatory reporting of “suspicion of abuse.” The duty to be law-abiding and the duty to be truthful, even about clinical impressions based partly on evidence and partly on intuition, compels the caregiver to explore the situation in order to substantiate or discard the hypothesis about the patient and then to act accordingly.

There is one interesting exception in which it has been agreed by the profession of medicine that “deception” has been viewed as ethically acceptable: the publication of case studies in which identification of the patient would be possible unless certain historical “details” are obscured in the presentation. Although it is optimal to obtain prospective consent for writing up a case, this may not always be possible. In such situations, it is within the traditions of clinical medicine to present cases in a manner that disguises the patient appropriately. This does not give permission for all types of inaccuracies, however. In composing these case studies, it is incumbent upon clinicians to use a minimum of “false” information in protecting the patient’s identity and to alter no fundamental features of the case such that the clinical teaching point is itself based on an inaccuracy. For example, one should not claim that an adult patient is a child in a published report in order to characterize a childhood syndrome. Similarly, one should not present “imagined” therapeutic processes and outcomes as factual when describing a novel patient care situation in a case report. However, stating that the patient grew up in “a small family in the Midwest” when he actually grew up in a large family in the Pacific Northwest, when these factors are not relevant to the clinical situation but can help protect the patient, is, in this special circumstance, viewed as ethically acceptable. When developing a publication of this type, authors ideally will seek advice and collaboration from others in making judgments about how much of the truth to reveal, how much to omit, and what kinds of purposefully inaccurate information to include so as to adequately disguise and protect the patient. In many circumstances, it may also be appropriate for the authors to acknowledge explicitly their efforts to guard the identity of the patient through such strategies in the published report. The fact that some information in the case study is altered must be noted for the editors at the time of submission as well.

A 17-year-old has been in psychotherapy for 2 years. He has generally been doing well in school and at home but continues to have irritable periods and occasional outbursts of anger that are frightening to him and to his family. His mother calls the therapist and wants to know, “What is going on with my son?”

A 22-year-old woman was physically assaulted by a stranger at a party. She is evaluated in an emergency room, and the physician becomes concerned about alcohol use and depressed mood as reported by the patient over the previous year. Her father, who holds a very high-ranking position within the community hospital, asks that the physician not document these concerns, or the circumstances of the assault, in the patient’s chart.

An employer calls a community psychiatrist and verbally requests information about a worker’s mental health and addiction history. The employer tells the clinician that he is entitled to this information because “the company pays” for the patient’s insurance, and the patient “has to be in good health to do his job.”

A 34-year-old mother of three small children sees a psychiatrist for panic attacks that began after she was in a car accident at age 17 years. She has been psychiatrically stable for years and requires only semiannual “checkups.” At her most recent visit, she tearfully confides in her psychiatrist, “I’m worried about my temper with my kids. I feel like a terrible mother, and I can’t keep myself from spanking them.”

(Reprinted with permission from Roberts LW, Dyer AR: Confidentiality and truth telling, in Concise Guide to Ethics in Mental Health Care. Arlington, VA, American Psychiatric Publishing, 2004, Chap 7)