Confidentiality of Health Information Postmortem

Medical professionals have long considered confidentiality of patients' medical information of paramount concern. The patient-physician relationship is, in large part, based on a trust that the information obtained within the relationship will remain confidential. However, confidentiality protections are not absolute, and a variety of exceptions are based on individual and public health concerns. Recently, attention has focused on the limits of the physician's duty to preserve confidentiality after a patient's death. The Proposed Rule of the Department of Health and Human Services on “Standards for Privacy of Individually Identifiable Health Information” recommends that privacy protections of medical information cease 2 years after death.¹

The Council on Ethical and Judicial Affairs (CEJA) of the American Medical Association offers the following report to identify limitations to confidentiality of medical information postmortem and situations in which physicians may disclose relevant information to third parties. For the purposes of this discussion, postmortem medical information refers to any information contained within a deceased patient's medical record, including information entered into the record after death. The report begins by discussing the premise and scope of confidentiality and then outlines factors physicians should consider in determining whether they may disclose information postmortem.

There are a number of bases for protecting confidentiality of medical information. One basis is the inherent value of privacy. The value of privacy derives, in part, from the tendency to fashion one's own identity and to control how much information about oneself to conceal or reveal to others.² Another more commonly cited rationale is the practical benefit of maintaining patient confidences. Confidentiality protections help assure patients that they can entrust to their physicians private information that is important to the provision of care. However, patient confidentiality is not absolute. Society has a legitimate interest in permitting (and sometimes promoting) breaches of confidentiality. For instance, physicians are required to report certain communicable diseases, gunshot or other wounds, and evidence of child abuse or neglect.³ In “Opinion 5.05, Confidentiality,” from the American Medical Association's Code of Medical Ethics, the Council acknowledges the potential for limited disclosure: “[t]he obligation to safeguard patient confidences is subject to certain exceptions which are ethically and legally justified because of overriding social considerations.” ⁴ The extent to which physicians have an obligation to maintain confidentiality of medical information may be superseded by other interests and concerns.

The inherent value of privacy and the practical benefits of maintaining confidentiality for the living also provide a foundation for protecting medical information postmortem. In contemporary US society, some individual interests survive death. For example, the practice of honoring wills functions to respect the interests of the deceased in controlling the distribution of property. Likewise, protecting confidentiality after death functions to respect the former interests of the deceased in controlling personal health information.

The obligations a physician may have to a deceased patient are less clear because privacy and confidentiality traditionally focus on a living patient's control over information. One might argue that disclosing information postmortem is of little consequence because the dead cannot be harmed, nor can they have any interest in confidentiality.⁵ But this stance ignores the potential harm to deceased patients' identity with respect to their legacy. Cicero wrote: “The life of the dead consists in being present in the minds of the living.” ⁶ Similarly, those who were close to the deceased hold interests in preserving the memory of their loved ones.

In addition to the inherent value of privacy, protecting confidentiality postmortem may also have practical benefits. To expand on the above example, the practice of honoring wills promotes both the interests of the deceased and also the interests of the living. People make wills on the assumption that their wishes will be implemented after their death. In other words, people have a current interest in ensuring that wills, in general, are enforced. Similarly, maintaining confidentiality of medical information postmortem assures living patients that the information they impart to their physician will not be disclosed after death. Disclosure of such information on a regular basis may weaken the institution of confidentiality as well as public trust in physicians.

Thus, the reasons for preserving confidentiality of health information for living patients seem to apply postmortem as well. However, the inability to obtain consent for disclosure from the deceased may influence the degree to which such information should be kept confidential in the face of conflicting interests. Consent, although a useful safeguard for living patients, is hardly helpful in this context. Therefore, we must fashion confidentiality protections that are not unduly restrictive.

One possibility is to borrow from the concept of surrogate decision making. In cases in which a patient receiving life-sustaining treatment loses decision-making capacity, the Council suggests that decisions be made by a surrogate decision maker. In the absence of a designated proxy (eg, through an advance directive), the patient's family should become the surrogate decision maker. If there is no person closely associated with the patient, but there are persons who both care about and have sufficient knowledge of the patient, then such persons may be appropriate surrogates.⁷ In this context, the surrogate example provides a useful template for identifying a proxy for deceased patients.

Although this example is helpful in some cases, a variety of problems arise, similar to problems that exist for living patients. These include the difficulty of ascertaining an appropriate decision maker, if one has not already been designated, as well as elucidating the patient's preferences. However, there is an important difference in the nature of the decision to be made by the surrogate of a living patient and the surrogate of a deceased patient. In the former case, decisions focus on treatments the patient would have chosen.⁸ In the medical context, patient autonomy is an expression of choice among various potential therapeutic benefits. In the latter case, decisions should reflect how the decedent would have wanted to control his or her lasting identity in general, and his or her health information in particular. Although this also constitutes an expression of autonomy, there are no therapeutic benefits that will come to rest with the deceased, regardless of a decision made by a surrogate. Simply articulating an individual's attitudes and values may be adequate to infer treatment decisions but may be inadequate in determining how to protect an individual's lasting identity.⁹ If the decision maker associates the choice to release private medical information with how that patient would construct his or her identity, it might be useful to think of the decision in the context of the patient's life story. Rather than use the substituted judgment standard to predict what the patient would have decided if living, some find it helpful to consider the patient's life story and how well a particular option fits with the patient's lasting story or identity.¹⁰

In the United States, the protection of confidential information postmortem varies between states. For example, in a number of states, autopsy reports performed under the auspices of a medical examiner become part of the public record.¹¹ In these cases, state Freedom of Information Acts commonly require that public records be available to anyone who wants them unless an exception applies. Exceptions restricting public access to medical information may apply across the board, as in Massachusetts, or may depend on a court to balance privacy interests against public interest in the disclosure of those records, as in New York.¹²

The American Medical Association's general policy regarding disclosure states: “Conflicts between a patient's right to privacy and a third party's need to know should be resolved in favor of patient privacy, except where that would result in serious health hazard or harm to the patient or others.” ¹³ Clearly, confidentiality protections postmortem would not be more stringent than those in place during a patient's life. These protections, at their strongest, would be equal to those for living patients. Specific to deceased patients, “Opinion 5.057, Confidentiality of HIV Status on Autopsy Reports,” from the Code of Medical Ethics, notes that in the absence of law, physicians should ”… fulfill ethical obligations to notify endangered third parties (e.g., identified sexual or needle-sharing partners).” ¹⁴ Thus, in this narrow case the Council recognizes a permissive notion of disclosure of confidential information postmortem and a possible obligation to warn at-risk individuals.

When deciding whether it is permissible to disclose medical information postmortem, one should weigh the interests in preserving confidentiality against the interests in disclosing the information. In most cases this determination will be based on ethical or legal criteria similar to those used to make judgments about the release of confidential information for living patients.¹⁵ We will assume that if information about a living patient may be disclosed ethically, the same information may likewise be disclosed after that patient has died.

There are at least 2 sets of circumstances in which information concerning a deceased patient might be sought. The first is when disclosure would provide information directly pertinent to the health of a particular individual. The second, which will be dealt with in the next section, is when the information sought would be used for research, education, or other purposes not directly affecting a particular individual's health.

The familial nature of genetic information creates the potential of gaining genetic information about blood relatives of the deceased. Similarly, information relating to an infectious disease may be of significant use to at-risk third parties. In the case of living patients, the treating physician can discuss disclosure with the affected patient or encourage the patient to avoid endangering others.¹⁶ Clearly, encouraging disclosure or avoidance is no longer an option postmortem. In this context, physicians considering disclosure should examine whether the potential for harm is likely to occur,^17,18 if the at-risk individual is identifiable,^19,20 and if disclosure is likely to be of benefit to the at-risk individual.²¹ When there is a threat to the public health, or when legally required to do so, physicians should disclose only necessary information to the appropriate authorities.

The basis for disclosing information postmortem is to protect at-risk third parties from conditions of which they may not be aware and to offer them the opportunity to be treated. However, it is important to recognize that in nearly all instances, a deceased patient's medical information cannot lead to the diagnosis of a living individual, only to a probability of developing a specific health problem. The physician therefore should explain the nature of the information, leaving the decision for further testing up to the individual.²² Finally, in the rare instances when prior to the death a patient has stated explicitly that certain information should not be disclosed, the above 3 criteria should determine whether information may be disclosed. In all instances, physicians should disclose only the information that is necessary to warn interested individuals.

Besides providing medical benefit to particular individuals, there are many other reasons why medical information may be sought after a patient's death. These range from research to education to such things as public interest in biographical data. When applicable, confidentiality should be maintained to the greatest possible degree. Thus, for many research and educational purposes, individual identifiers can be removed from the information and it may be used as necessary postmortem.

In other cases, such as biographical studies, identifiers are crucial and disclosure is public. In these instances, physicians should consider any statement regarding disclosure that was made prior to the patient's death. In the absence of such guidance, the impact disclosure may have on the reputation of the deceased patient is an important consideration. Accordingly, those who hold an interest in preserving a certain memory of the deceased (eg, family members) should be involved in these decisions to disclose information. In all cases, physicians should be sure that personal gain for himself or herself is not the primary motivation for disclosure.

Finally, when consent is required to disclose information concerning a deceased patient (eg, autopsy results), what is to be disclosed is the decision of the individual granting consent to disclose. Otherwise, only limited information should be disclosed.

A number of concerns arise when assessing the appropriateness of disclosure of medical information postmortem. In all cases, physicians should consider whether harm is likely to occur in the absence of disclosure, whether an at-risk individual is identifiable, and whether the disclosure is likely to be of benefit to the at-risk third party. Furthermore, any statement regarding postmortem disclosure of information made by the patient prior to death, the impact disclosure may have on the patient's lasting reputation, and whether personal gain is a motivating factor for disclosure should also be taken into account. Actual disclosure of medical information should be responsive to both the needs of surviving individuals and the lasting interests of the deceased, reflecting the nature of the information being provided. Only the information that is necessary to adequately inform or warn third parties or public health authorities should be disclosed to those persons. To facilitate the advancement of medicine, a deceased patient's health information, with individual identifiers removed, may be used for educational and research purposes.

For the foregoing reasons, the Council recommends the following guidelines.

The Council on Ethical and Judicial Affairs (CEJA) of the American Medical Association (AMA) formulates ethics policies for the medical profession through its interpretations of the AMA's Principles of Medical Ethics. Members of CEJA at the time this report was adopted were Herbert Rakatansky, MD, Providence, RI (Chair); Frank A. Riddick, Jr, MD, New Orleans, La (Vice-Chair); Michael S. Goldrich, MD, Highland Park, NJ; Leonard J. Morse, MD, Worcester, Mass; John M. O'Bannon, III, MD, Richmond, Va; Priscilla Ray, MD, Houston, Tex; Robert M. Sade, MD, Charleston, SC; Monique A. Spillman, MD, PhD, Boston, Mass (Resident Member); and Matthew Weiss, Chicago, Ill (Student Member). Staff for CEJA included Audiey Kao, MD, PhD (Acting Vice President), Ethics Standards Group, AMA; Karine Morin, LLM (Council Secretary); Jessica Berg, JD (former Council Secretary); Andrew Maixner (Council Staff Associate and Staff Author); and Sam Seiden (Council Staff Associate).