Medical Privacy and the Public's Right to Vote: What Presidential Candidates Should Disclose

Abstract

We argue that while presidential candidates have the right to medical privacy, the public nature and importance of the presidency generates a moral requirement that candidates waive those rights in certain circumstances. Specifically, candidates are required to disclose information about medical conditions that are likely to seriously undermine their ability to fulfill what we call the “core functions” of the office of the presidency. This requirement exists because (1) people have the right to be governed only with their consent, (2) people's consent is meaningful only when they have access to information necessary for making informed voting decisions, (3) such information is necessary for making informed voting decisions, and (4) there are no countervailing reasons sufficiently strong to override this right. We also investigate alternative mechanisms for legally encouraging or requiring disclosure. Protecting the public's right to this information is of particular importance because of the documented history of deception and secrecy regarding the health of presidents and presidential candidates.

Keywords:

• democracy
• medical privacy
• presidential candidates
• right to know
• right to vote

I. INTRODUCTION

Presidential candidates, like everyone else, have a right to medical privacy. For most people, this right to medical privacy altogether precludes the public from viewing their medical records. However, in virtue of the very public role of the president, the idea that the public may be kept in the dark about the health of presidential candidates is untenable. Our purpose in this article is to make it clear that candidates are morally required to waive their right to medical privacy concerning a very specific set of medical conditions. Although others have asserted a moral duty to disclose (See, e.g., Annas, 2000), the literature contains very little discussion of the basis for that requirement. We argue that it is based on the same deep democratic principle that supports the public's right to vote, namely, that those who govern do so only with the consent of the governed. Concerns about the medical privacy of candidates must be subordinated to that democratic principle.

The health of presidential candidates is of particular concern, both because of the unique position of the president, and the inordinate pressures inherent in the office that can negatively impact physical and psychological health. ¹ The issue is not merely an academic concern, in light of a documented history of deception and secrecy regarding the health of presidents and presidential candidates (Crispell & Gomez, 1988; Gibert, 1992; Ferrell, 1992; Tsongas, 1992).

While campaigning for his first election, Woodrow Wilson hid the fact that he suffered at least three minor strokes, probably due to uncontrolled hypertension, which led to his massive stroke in 1919 (Ferrell, 1992, pp. 12–13). At the time, the etiology of strokes and their relationship to high blood pressure were not well understood. Nonetheless, the diagnosis was clear to his treating physicians and was deliberately kept from the public, which was told that he had merely suffered “a nervous breakdown, indigestion, and a depleted system” (p. 16). When his vice president, Thomas Marshall, refused to take over the presidency, Wilson's wife performed the presidential duties for the remainder of his term.

Franklin D. Roosevelt's complications from uncontrolled hypertension were also covered up during his presidency and his final re-election campaign in 1944 (Crispell & Gomez, 1988, pp. 114–120). Roosevelt had significant leg weakness due to polio, and after its onset used a wheelchair or leg braces (p. 86, et sub.); however, he stood when he orated in such a way as to convey an image of physical robustness, and there was a tacit agreement among the press, photographers, and even cartoonists not to depict Roosevelt's condition (pp. 83–84). He died from a cerebral hemorrhage in April of 1945, just months into his fourth term (pp. 152–154).

Dwight Eisenhower experienced a serious myocardial infarction during his first term as president, and it was very likely that he would not survive a second term. During Eisenhower's re-election campaign, cardiologist Paul Dudley White declared to the public that the president probably had five to ten years of active life left, despite statistical evidence that clearly indicated otherwise (Ferrell, 1992, pp. 94–110, esp. 109–110). Eisenhower's campaign was successful and despite the odds against it, he went on to complete a second term.

Paul Tsongas received a bone-marrow transplant to treat his non-Hodgkin's lymphoma in 1986. Even though follow-up tests disclosed recurrent lymphoma (subsequently treated in 1987), Tsongas and his physicians continued to report during his 1992 presidential campaign that he had been “cancer-free” since the bone-marrow transplant, a deliberate deception. And unbeknownst to the public, he suffered a relapse during the campaign. Tsongas disclosed the details of the situation only after he had suspended his campaign (Altman, 1993; Tsongas, 1992). Had he been elected, his performance would likely have been seriously impaired by the cancer treatment, and he would have died in office.

Clearly, past U.S. presidents and candidates have successfully hidden serious medical conditions during their campaigns, conditions with significant consequences for their ability to perform their presidential duties. In this article, we argue that voters have a right to this kind of medical information, and that such deception cannot be justified in view of that right. We begin by discussing the right to vote and arguing that the right to vote supports the public's right to have information necessary for making informed voting decisions. Next, we propose a standard which we argue strikes the right balance between the public's right to information, the candidates' interest in medical privacy, and everyone's interest that any information disclosed not be irrelevant or misused. We then respond to several objections to disclosure of medical information that are raised in the literature. We conclude by discussing the institutional mechanisms by which the public's right to know could be implemented and the implications of our view for journalists, candidates' physicians, and the candidates themselves.

II THE PUBLIC'S RIGHT TO VOTE AND ITS IMPLICATIONS FOR DISCLOSURE

On any plausible conception of democratic governance, citizens have the right to vote, and it is therefore morally problematic if voters do not have access to important facts about matters they vote on. ² For example, it would be problematic if citizens were unable to learn about candidates' views on important policy matters before they choose between them. There are numerous democratic principles that underwrite the right to vote: that those who govern should do so with the consent of the governed; that the government should represent the people; and that the people should be able to hold the government accountable. In turn, these principles underwrite ancillary rights: the right to physical access to the polls; the right to information about candidates' voting records; and the right to information about campaign financing. The moral importance of access to candidates' health information arises out of one of these democratic principles, that those who govern do so with the consent of the governed.

The right to vote secures the consent of the governed in two ways. First, it involves citizens in their governance in a way that makes them willing to submit to the rule of law and that justifies that willingness. More importantly, though, electing government officials by vote is itself a way in which the community as a whole gives consent to be governed by those elected. Although it is difficult to state precisely the way in which voting (or democracy more generally) secures the consent of the governed, or indeed the sense of “consent” in which those in a democracy do consent to be governed, the idea that voting is essential to government by consent is a familiar one.

It is also a commonplace that the meaningfulness of any sort of consent depends on the availability of information to the decision maker. This can be construed in at least two mutually compatible ways.

First, the meaningfulness of consent depends on the decision maker's having access to valued information. It is worth noting that people often want information that turns out not to impact their final decision. For example, many patients want to know whether they are being used for the training of medical students, even when they have no objections to being so used. Similarly, information about possible outcomes of a decision can be meaningful even if those outcomes do not occur. So even if the patient ends up surviving, consent is not fully informed if she would have avoided the procedure had she been informed of its risk of death. Thus, information can be relevant to the meaningfulness of consent even if it does not impact one's final decision and even if the outcomes for which the information is relevant do not occur.

Second, the moral force of consent increases as more relevant information becomes available to a decision maker. So for our purposes, where voters desire relevant information about a decision, but where that information is unavailable (either because it is undocumented or because it is documented but undisclosed), the quality of that consent decreases. While meaningfulness of consent admits of degrees, we can say that consent is only valid where that consent is properly informed, leaving open for the moment just what information is necessary for consent to count as informed. ³

But if voters have a right that their valid consent be obtained, and if consent is not valid unless it is informed, it follows that voters have a right to the information necessary to make an informed voting decision. A fortiori, if certain kinds of medical information are necessary for making informed voting decisions, voters have a right to that medical information. This explains why the historical examples we mentioned above are problematic. The Wilson case is clearest in this regard: his illness resulted in the effective disenfranchisement of the entire American people who had no idea that their vote for Wilson would give the power to make presidential decisions to his wife.

Tsongas's unsuccessful 1992 presidential campaign also clearly highlights how complete confidentiality could undermine voters' rights to make informed voting decisions. If Tsongas had been elected, it was unlikely that he would have lived through his term, due to his recurrent lymphoma. Clearly, those voters who would have voted for Tsongas had he run without disclosing his condition, but who would not have voted for him if they knew about his recurrent cancer, would not have been making an informed voting decision. A second group might have voted for him even had they known about his recurrent cancer; even so, they would likely have considered access to that information important in their deliberations. Thus, their vote would not have been fully consensual in light of the fact that they were deprived of that information, and their vote would still have not been fully consensual even if Tsongas had been elected and had not suffered any complications of his cancer.

Similarly, Eisenhower's failure to disclose his medical conditions illustrates a situation in which the voters were denied the ability to make an informed voting decision even though the medical condition did not have the expected outcome.

Finally, Roosevelt's 1944 re-election campaign illustrates a situation in which withholding information from the voters can render their decision uninformed even though, because of Roosevelt's overwhelming popularity, the information probably would not have altered the outcome of the election.

Now, the mere fact that a person has a right to certain information does not by itself imply that those who possess the information should disclose it, for it may be that there are other, countervailing considerations of sufficient importance to override that right. This is acknowledged in the discussion of informed consent in the medical ethics literature by distinguishing between the information a patient needs to make an autonomous decision about his or her healthcare, and the information a doctor is morally required to provide, all things considered. Consider the case of a patient who is deeply concerned about the success rates of his doctor and hospital to the point where he wants detailed statistics on how the doctor's and hospital's patients have fared over the years. This may well be material to this particular patient's decision, and thus its provision may well be included in the patient's right to informed consent, but the economic and logistical costs of collecting such information may be prohibitive. ⁴

Similarly, in the case of voting, there is a distinction between the information to which voters have a (prima facie) right and the information that candidates are morally required, all things considered, to disclose. The right to information necessary to make informed voting decisions must be weighed against what is fair and reasonable within the existing political system. For example, suppose that candidate Mary Smith received medical care that obviously shows that she was a rape victim. Even though this may be relevant to some citizens' voting decisions (either because they think it presages certain policies or because they think it reflects something about Smith's personal disposition and lifestyle), Mary Smith is clearly not morally required to disclose such information.

It is worth pausing here to be explicit about our analogy between government by consent and consent to medical care. Patients and citizens are analogous in that they are both in a position to consent to the actions of decision makers, who are thereby authorized to exercise power over them, and in each case their right to informed consent demands that they have adequate information. Thus, citizens and their elected representatives on the one hand, and patients and their healthcare providers on the other, sit in similar relations to each other vis-à-vis consent. In the case of presidential candidates' medical information, this analogy is complicated by the fact that it is the candidate's information in his role as a patient that is relevant to citizens' consent to be governed. The proper analogue of presidential medical information in the patient-healthcare provider model, therefore, would be medical information about a healthcare provider that is important for a patient's treatment. A patient's right to informed consent, for example, would entitle them to information about a surgeon whose hands are numb or an anesthesiologist with narcolepsy because that information is material to their treatment decisions. ⁵

The current legal system allows candidates to maintain complete medical confidentiality, if they desire; voters are thus denied any mechanism for pressing their right to even the most important medical information. But the Mary Smith example shows that merely asserting that voters' informational rights must be taken more seriously is overly simplistic. Rather, to determine whether a candidate is under a moral requirement to disclose information, what is needed is a principled distinction that strikes an apt balance between voters' informational rights, candidates' privacy, and other democratic goals and values. To be clear, where we say “moral requirement,” we mean an undefeated moral duty, such that an agent ought, all things considered, to comply with that duty.

So, what information is a candidate morally required to disclose? We propose that candidates are morally required to disclose information about any medical conditions that are likely to seriously undermine the candidate's ability to fulfill what we will call the core functions of the office. “Likely” and “seriously” are of course vague, but any plausible view will have to be qualified in some such way, since any plausible view will take as relevant the severity and likelihood that the medical condition will impact the candidate's abilities. ⁶ The factors that are relevant to deciding in particular cases whether a certain piece of medical information must be disclosed will vary in degree, and apart from the extreme cases, it will be vague where the precise balance lies, thus necessitating the exercise of judgment on a case-by-case basis. We explicate our use of these terms below, and provide some guiding examples in Section III.

By “core functions of office” we mean those functions that, on any reasonable view, are necessary to perform the job of the presidency. So while citizens have vastly different views of the particular functions of the presidency, an adherent of any of these views would agree on what the core functions of office are. To take some extreme examples: on any reasonable view of the president's job, the president must be alive and conscious to perform it. Similarly, on any reasonable view, serious mental illness will likely impair the president's ability to perform the functions of office. So it seems very plausible to say that a president must be alive, conscious, and competent in order to perform the core functions of the office. By contrast, although on some reasonable views, intemperance might impair the president's ability to perform the functions of office, on other reasonable views, intemperance is unlikely to do so. Thus, being temperate is not required to fulfill the core functions of the office.

The requirement we propose is quite strict, adhering only to these uncontroversial claims about the abilities the president needs. It is limited to a narrow class of information (i.e., what is likely to affect a candidate's ability to fulfill the core functions of office) and sets a high relevance threshold (i.e., what is likely to have a serious effect). Such a strict requirement has several advantages over more expansive disclosure requirements. ⁷

First, by limiting disclosure to information relevant only to the core functions of office, our view does not require candidates to disclose information that is irrelevant or only controversially relevant to their ability to perform as president. Thus, while our view would have required Tsongas to disclose that his cancer had relapsed, it would not require Mary Smith to disclose information about having been raped. Thus, while requiring disclosure of potentially damaging but uncontroversially relevant information, our view avoids requiring disclosure of potentially damaging but at best controversially relevant information.

Second, by limiting disclosure to information about conditions that are likely to have serious effects, our view does not require disclosure of all medical conditions that might have some bearing, no matter how small, on the candidate's ability to perform the core functions of the office. Just as a patient's right to informed consent does not require the doctor to disclose absolutely every risk and benefit of a treatment, no matter how minor or remote, the fact that a candidate has sensitive skin or minor asthma could conceivably affect his ability to fulfill the functions of office, but would not need to be disclosed under the standard we propose. This ensures that information about important risks and benefits is not lost amid a sea of mostly useless information, and it provides a procedural safeguard against a slippery slope towards disclosing inappropriate medical information.

In addition to the advantages already mentioned, limiting disclosure in both of these ways protects candidates' important privacy interests and protects the political process against those who might hijack political deliberation with specious arguments about relatively trivial medical matters. And, as we will discuss in more detail in Section IV, it does so while giving due weight to the rights and interests of voters, unlike a view that permits complete confidentiality. ⁸

III EXAMPLES OF CONDITIONS REQUIRING DISCLOSURE

Our view has several specific implications regarding the kinds of conditions that candidates must disclose, including the following: known medical conditions that would give the candidate a life expectancy of five years or less; medical conditions that would significantly impair the candidate's judgment or behavior while in office; history of past illness likely to recur or cause complications later in life; and mental illness likely to result in significant cognitive impairment while in office. The examples that follow serve to illustrate how to apply our standard.

A Known Medical Conditions

As noted, Tsongas concealed his relapse of non-Hodgkin's lymphoma from voters and campaigned in the 1992 Democratic primary. He died before the end of what would have been his first term. In our view he would have been required to disclose his condition, not merely because it turned out to be fatal, but because it was known at the time to have a high probability of being fatal for anyone who had it. More generally, our view requires candidates to reveal any medical condition that gives them a life expectancy of five years or less. This is because voters consent to be governed one term at a time, and five years is roughly one term, plus the time it takes to campaign for the office. One might object that prognostication is inexact, and that therefore this standard is unworkable. George Annas, for example, says that “physicians cannot predict with any certainty how a candidate's medical condition will affect his or her ability to perform politically” (Annas, 1995, p. 947). But there are in fact a number of conditions for which life expectancy can be predicted with fair precision and certainty. For example, metastatic cancer, cardiomyopathy, and advanced AIDS all carry defined five-year survival rates (Solomon et al., 2001).

In our view, candidates are not required to disclose only conditions that are fatal, for there are many non-fatal medical conditions that are likely to affect a candidate's judgment or behavior so substantially as to compromise their ability to fulfill the core functions of office. For example, candidates would also have to disclose uncontrolled epilepsy, a history of undiagnosed syncope, significant ventricular arrhythmias, and chemotherapy (given its potential toxicity).

Despite the fact that many conditions meet our standard, most medical conditions would not require disclosure. Rudy Giuliani's early prostate carcinoma, Bill Bradley's atrial fibrillation, and John McCain's early-stage melanoma were unlikely to alter their longevity or cognition. Although these politicians voluntarily revealed their conditions (Purnick, 2000), our view would not have required disclosure. More directly related to our topic, though, is Vice President Dick Cheney's heart condition. Cheney has had several heart attacks and has received various cardiac treatments in recent years. Cheney did not release his full medical records during the 2000 campaign, and the issue did not attract much attention during the 2004 campaign (Prothero, 2000; Schmitt, 2001). ⁹ The question arises whether our view would have required disclosure. There are a couple of reasons to think that it would not. First, Cheney was only aspiring to the vice presidency, and the likelihood of vice presidential incapacitation precipitating the sort of disruption that grounds the disclosure obligation is much smaller than the likelihood of presidential incapacitation doing so. Moreover, his heart condition does not rise to the level of a high likelihood of death within 5 years, given the many technological advances in treating heart disease.

In addition to the historical cases we have discussed, the popular television drama The West Wing illustrates that this is a concern to the public. In this show, President Bartlett concealed his previously diagnosed mild multiple sclerosis during his campaign. When he suffered a relapse during office, the characters in the show hotly debated whether Bartlett should have informed the public of his condition during his campaign; Aaron Sorkin, the primary writer for the show, decided to have Congress issue a censure of him (Sorkin, 2001). Because mild multiple sclerosis is unlikely to affect five-year longevity, cognition, or competence, our view would not require disclosure. The show raises other interesting issues—Bartlett must work to regain the trust of his advisers and staff, Bartlett's wife, who happens to be a physician, colludes in the cover-up—but these alone do not suffice to meet our standard for required disclosure. So even some important medical conditions would not, in our view, require disclosure.

B Past History of Illness

Whether candidates are required to disclose past illnesses that might resurface under the stress of the presidency depends on the likelihood of recurrence and the seriousness and likelihood of anticipated complications. Many childhood cancers are completely curable and very unlikely to recur, but others, such as Hodgkin's disease, are associated with an increased risk of second malignancies and leukemia later in life. Whether a candidate will have to disclose such a history, then, will depend on whether it is likely to seriously impair a president's performance during the term for which he or she is running. Each case is unique, and depends on such factors as the stage of the disease, the time elapsed since treatment, and the kind of treatment used. We propose one procedure below for adjudicating such issues on a case-by-case basis.

C Mental Illness

Several presidents and presidential candidates have suffered from mental illness, thus raising the question of whether candidates are morally required to disclose information about current or past mental illness.

In 1972, Thomas Eagleton resigned as George McGovern's running mate after it was publicized that he had had three psychiatric hospitalizations for major depression, two of which required electroconvulsive therapy. Although his political career was unblemished by any depression-related incidents, the press conjured scenarios of a compromised leader in charge of a nuclear arsenal. McGovern initially supported Eagleton “1000%,” but requested his resignation after their standing in the polls plummeted (Crispell & Gomez, 1988, pp. 5–7). James Reston summarized what many Americans felt:

Maybe the Republic can bear this human compassion in the Supreme Court and the Congress—even though there it is highly questionable—but at the level of the Presidency and the Vice Presidency in this age of atomic weapons abroad and human violence and political assassinations at home, the present system is wildly out of date. (Reston, 1972)

Many historians attribute Calvin Coolidge's weakened presidency to unrelenting depression and self-blame over the death of his youngest son the year Coolidge was elected (Gilbert, 1992, p. 41). More recently, it was revealed that Nixon saw a psychiatrist and was prescribed phenytoin for his mood swings while in office (Clymer, 2000b). He was diagnosed with “neurotic symptoms: anxiety and sleeplessness” (Clymer, 2000a). His behavior during the Watergate scandal clearly suggested clinical paranoia and doubtless affected his ability to govern the country.

Effective medical treatment for depression was developed in the late 1960s, with the advent of tricyclic antidepressants and monoamine oxidase (MAO) inhibitors. Electroconvulsive therapy, while effective in many cases, left many patients with permanent memory deficits, definitely a handicap for a potential U.S. president. However, mental illness, even depression and anxiety, still carries a significant stigma and is ordinarily protected by ironclad confidentiality.

Should a history of depression be disclosed? Advocates for the mentally ill have worked hard to destigmatize mental illness and ensure that it not be used to discriminate against job applicants. However, in some cases the effects of depression—which may include trouble concentrating, impaired memory, and insomnia—can be severe enough to undermine one's ability to adequately perform the core responsibilities of a position like the presidency. Moreover, the extreme stressfulness of the U.S. presidency, which requires shrewd judgment and decisional capacity, but which frequently compromises sleep, could easily trigger a recurrence of depression and could be especially conducive to cognitive impairment.

Eagleton, when asked why he had not been more forthcoming initially, replied that he considered his condition no more serious than a broken leg or arm (Crispell & Gomez, 1988, p. 4). This response appears disingenuous in the face of the severity of his illness, which required hospitalization and electroconvulsive therapy, although it is impossible to say whether Eagleton would have required hospitalization and ECT if he had had access to modern psychopharmacology with Prozac (fluoxitene) or one of the many effective drugs for depression.

It is true that requiring disclosure of such conditions could lead to decreased chances of success for otherwise well-qualified candidates; however, a prior episode of depression does increase the likelihood of recurrence (Thase, 1990, pp. 51–57), and this clearly seems relevant to voters' decisions. Depending on the best available evidence as to the likelihood and severity of recurrence, our view might well require disclosure of such a condition. In Eagleton's case, it almost certainly would have.

D Irrelevant and Inappropriately Relevant Information

Although information about certain medical conditions is necessary for making informed voting decisions, other kinds of information, though considered relevant by some, need not be disclosed since they do not bear on a candidate's ability to carry out the core duties of office. Examples include information about sexual preference, past drug use, and extramarital affairs. Disclosure of such stigmatizing information severely compromises a candidate's privacy without justification and in our view would not be required. This parallels the case of private employers, who cannot require disclosure of medical information that is irrelevant to how well an applicant can perform a job. ¹⁰

It might be objected that disclosing such information would be required in our view, at least in situations in which the public or Congress were likely to discover the information and react so vehemently that the president would no longer be able to do anything other than spend his time on damage control. Such information might be called “inappropriately relevant information,” since it is information that is in itself irrelevant to the candidate's job performance, but can come to be relevant when mistakenly perceived to be so by others. This phenomenon is exemplified by the Clinton/Lewinsky scandal.

There are several reasons why disclosure of such information would not be required in our view, even though it is, in a way, relevant. First, these are conditions whose effects upon a candidate's ability to perform the core functions of office depend upon their being disclosed, and the likelihood of accidental disclosure is not amenable to prediction in the same way as, say, the effects of non-Hodgkin's lymphoma. Second, even when these conditions are accidentally disclosed, the impact of disclosure is also not as predictable as the strictly medical conditions discussed above. Moreover, even if disclosure does come to consume Congress and the public's attention, it is still not clear that it is likely to seriously undermine the president's ability to fulfill the core functions of the office. Scandals of all sorts seem to roll off candidates like water off a duck's back to sitting presidents (c.f. Reagan and the Iran/Contra affair, and the Clinton/Lewinsky scandal).

Finally, even in the unlikely case where such information is likely to be disclosed and would seriously undermine a candidate's ability to perform the core functions of office, it would still be inappropriate to require disclosure. Consider that in any other context, it would be unjustified to require disclosure of a person's private information merely on the grounds that the reactions of others will hinder that person's ability to perform his or her job. For example, suppose that an employer wants to collect information about potential employees' sexual orientation. The employer might argue that this information really is relevant to job performance because, were the employer to find out that an employee was homosexual, the employee would then be unable to work due to the resulting hostile work environment. But preemptive questions that contemplate an employee's ability to function only given people's attitudes are of a different kind than questions that would affect that person's ability to function regardless of anyone's perceptions. ¹¹ That kind of specious reasoning clearly does not justify an invasion of the employee's privacy. Thus, the general requirement to disclose information about any condition likely to seriously undermine a candidate's ability to perform the core functions of office must be qualified to exclude inappropriately relevant information.

IV THE CASE AGAINST DISCLOSURE REQUIREMENTS

There are a number of arguments that one could make in favor of thinking that candidates are morally permitted to maintain complete confidentiality. Here we respond to several of the most plausible arguments.

A The 25^th Amendment Makes Disclosure Unnecessary

The 25^th Amendment provides the constitutional mechanism for presidential succession, and information to this effect is readily available to the voters whose consent is required for governance. Thus, one might argue that because this mechanism picks out a particular person whom the populace votes on (viz., the vice president), their informed consent has been secured because they vote on a ticket of president and vice president, even without disclosure.

This argument is flawed, though, because there will always be important differences between presidential and vice presidential candidates, and in the absence of any evidence to the contrary, citizens expect that they will be governed by the one and not the other. Information that meets our standard is clearly relevant to the likelihood of presidential succession, and is therefore relevant for informed consent, even on the assumption that the transition to the vice president goes smoothly. ¹² Consider another surgery analogy. One might consent to surgery from a team with both a primary and a backup surgeon, and exchanging the primary for the backup surgeon because the primary becomes incapacitated would not render consent uninformed. However, if there are substantial differences between the two, if the patient expects that the primary surgeon will perform the procedure, and if there are circumstances that make it significantly more likely that such an exchange will occur, meaningful consent requires disclosure of those circumstances. Knowledge of such circumstances, when they exist, may well lead the patient to consider the qualifications of the backup surgeon more carefully.

To take a contemporary example: it seems plausible that Bush's incapacitation would not be as disruptive to policy making as the incapacitation of other presidents would have been, perhaps because Vice President Cheney plays a particularly active and able role in the Bush administration. (Compare with the possibility of Dan Quayle succeeding George H. W. Bush). But there are nonetheless significant differences between Cheney's views and Bush's views (e.g., on a constitutional amendment to ban gay marriage), and perhaps significant differences in the way they are viewed by other world leaders. So even if the 25^th Amendment did suffice to ensure a smooth transition in case of presidential incapacity, information about the likelihood of a transition is relevant to the voters' decisions because of such differences.

This is not to say the 25^th Amendment is irrelevant to our discussion of mandatory disclosure, since conditions that are likely to lead to the exercise of 25^th Amendment succession procedures are conditions that would clearly require disclosure. ¹³ The 25^th Amendment contemplates removal on the grounds that the president is unable to “discharge the powers and duties of office,” (U.S. Const. Amend. XXV, § 4) and it follows a fortiori that such conditions would undermine the president's ability to carry out the core functions of office, which is required by our standard. But there are conditions that meet our standard, but that are not likely to trigger the 25^th Amendment, specifically conditions that render the president significantly less able, but not unable, to fulfill the functions of office.

B Candidates Should Be Judged on Their Record of Service

Eugene McCarthy refused to disclose his medical records in the 1976 presidential campaign, stating that the president should be elected “on the basis of his or her record of service, of thought about the issues and programs to deal with them, and not on the basis of any private status such as that of patient” (Bloom, 1976, cited in Annas, 1995). But a moment's reflection shows that voting decisions should be based not just on the candidate's past record of service and views about issues and programs, but also on the candidate's ability to implement those views and programs were he or she to be elected. And plainly enough, there are many medical conditions which can affect that ability. Moreover, the potential disruption that could result from the president being incapacitated—regardless of what that incapacity does for the future of any particular policy or idea—is enough to make the candidate's health appropriate grounds for voters' decisions.

C Disclosure Requirements Would Violate Candidates' Medical Privacy

During his 1992 candidacy, Bill Clinton initially refused to disclose any specific medical information out of a desire to protect his privacy, thus expressing the widely shared concern that disclosure requirements somehow violate candidates' medical privacy (Altman, 1992). There are two such arguments that are worth considering.

First, it might be argued that a moral requirement to disclose is inconsistent with the right to medical privacy. After all, what is the right to medical privacy but a moral permission to withhold private medical information (except in extreme circumstances such as public health emergencies), and if candidates have permission to withhold such information, how can they be required to disclose it?

But the existence of a moral requirement to provide medical information is perfectly consistent with the right to medical privacy. This is because the right to medical privacy is a right against others not to disclose private medical information without consent, and the fact that others have a duty not to disclose a candidate's medical information is consistent with the candidates themselves being required to disclose that information. As is often noted, to say that someone has a right to do something is not to say that it is right for them to do that thing. ¹⁴

A candidate's right to privacy would be compromised if healthcare providers were legally required or legally permitted to disclose information without the candidate's consent. But even if one were tempted to try to legally institute mandatory disclosure requirements, such requirements would not legally compel or authorize anyone to disclose information without the consent of the candidates; rather, they would make consent a condition of candidacy. This is something we think is unproblematic in other contexts. For example, while employers may not access an employee's medical record without consent, they may require the employee to make relevant medical information available as a condition of employment. In sum, the right to medical privacy is not a good reason for thinking that there is no moral requirement to disclose, nor is it a good reason for thinking that there should not be a legal requirement to disclose.

Second, it might be argued that even if requiring disclosure of relevant information does not violate a candidate's right to medical privacy, there still remain important privacy interests that ought to be weighed in the balance while deciding whether someone is morally required to forego confidentiality. Someone with AIDS, for example, might not only be concerned about unconsented disclosure, they might also be interested in minimizing disclosure, period, with or without consent. People with chronic back problems have an interest in not foregoing confidentiality if their employers can use that information to avoid paying worker's compensation claims. And most people do not want others to know even innocuous medical information about them, since it isn't any of their business. All of these are important privacy interests that are independent of whether disclosure of the information occurs with or without consent.

Although these are important interests, they are justifiably compromised in many cases. An employer's interests in making an informed hiring decision about a potential employee overrides the potential employee's interests in privacy, and thus employers may require disclosure of medical information relevant to potential employees' expected job performance. Airline pilots, for example, are required to pass stringent medical and psychological examinations administered by FAA-approved physicians, every six months (Bureau of Labor Statistics, 2004). The same holds for presidential candidates: The voters' interests in making informed voting decisions override the candidate's interests in keeping private any information that meets our standard of relevance.

D Disclosure Requirements Will Deter Candidates

One might argue, as George Annas does, that the public interest is best served by making sure that potential presidential candidates are not deterred from seeking medical care for fear of its effect upon their candidacy (Annas, 1995, p. 5), and that otherwise qualified candidates are not deterred from running for office on the basis of their medical profile.

We agree that the public has an interest that fear of disclosure not deter candidates from seeking medical care, nor that it deter good candidates from running for office. The public's interest in candidates seeking medical care arises in two ways. First, the public has an interest in current candidates seeking healthcare in order that its future president be well cared for. But this interest does not conflict with our disclosure standard. Our view only requires disclosure of conditions that are likely to seriously undermine the candidate's ability to perform the core functions of the office, and the basis for that interest is an interest in the future president's ability to carry out the duties of office. If information about the candidate meets the standard for which we argue, disclosure would only be required when that ability is seriously questioned. As to the argument that disclosure requirements might deter otherwise good candidates from running for office, this requires saying that a particular candidate is a good choice for holding a four-year position, despite its being likely that he or she won't be able to carry out its duties for the full four years. So, although one could point to the effective year or so of Roosevelt's last term as an example of a candidate who was well-qualified despite his health problems, one would have to argue that having him in office for a year was a better choice than having some other candidate for all four years. While possible, this seems unlikely. More importantly, though, it is up to the public to decide whether a candidate's good qualities outweigh the possibility that their medical condition will obviate those qualities.

Second, the public has an interest that people who might become candidates receive medical care. Presidential candidates come from a fairly rarefied pool of people, such as senators, representatives, governors, and vice presidents, many of whom may have long-standing presidential aspirations. If they fear that medical care will result in disclosure of potentially damaging information, they might forego that care. If they do, this would be contrary to the public's interest in those people being healthy. This interest, too, is unlikely to be threatened by our standard. If the medical condition for which a potential candidate might seek care is minor, then it would be unlikely to meet our standard, and no disclosure would be required. And if the condition is major, such as lymphoma, then the costs of avoiding care may be enough to motivate the candidate to get care, and, at any rate, avoiding care might result in failing health, which would harm his or her chance of becoming president in its own right.

E Disclosure Requirements Will Infringe Voters' Rights

One might object to a moral requirement for disclosure on the grounds that compliance with that requirement would undermine some people's voting rights. For example, members of an advocacy group might seek to elect a candidate with a particular condition (or merely hope that one is fortuitously elected) because it will aid their cause. If candidates disclosed information about their conditions, this might harm the advocacy group's chances of electing their candidate of choice. Consider the case of hefty filing fees to get one's name on the ballot. The effect of such filing fees is that affluent groups will be able to place the candidate of their choice on the ballot, while poor groups will not be able to do so. The ability of poor groups to vote for their candidates of choice is thus diluted compared to the ability of affluent groups to vote for their candidates of choice. Such filing fees are morally unacceptable precisely because of that effect on some people's voting rights. In fact, the Supreme Court held that such fees are unconstitutional in Bullock v. Carter (1972), and their decision was surely morally correct (Gordon, 1976).

However, the analogous argument against our disclosure standard is unsound. The election of a candidate with a certain condition could aid the cause of an advocacy group either because the candidate would be an especially effective advocate or would provide a shining example of the lofty achievements possible for someone with that condition. But a candidate is likely to be a much more effective advocate by advertising, rather than hiding, the condition in question. Moreover, any benefits of post-election publicity would be marred by the public's realization that the candidate hid the information. And although a candidate's succumbing to a disease in office might arouse awareness and support, conditions which are likely to seriously undermine the president's ability to perform the core functions of the office are also likely to undermine the candidate's ability to be a successful advocate and a shining example. Thus, the interests of such an advocacy group do not conflict with our proposal.

It is also implausible to say that any requirement that makes it more difficult for a group's preferred candidate to be elected infringes that group's voting rights. There are any number of moral responsibilities incumbent upon a candidate that make it more difficult for him or her to be elected. Elements of a candidate's political career that are part of the public record are required to be accessible. Candidates ought to avoid lying or misleading the public about elements of their political platform. Clearly, these could disadvantage some groups' preferred candidates more than others.

And in rejecting such a requirement, the advocacy group would be trying to prevent other voters from having access to highly relevant information about the candidate, which runs contrary to democratic principles rather than preserving them. Disclosure of conditions that meet our standard will only provide voters with information highly relevant to the candidate's ability to function in the presidency, and so will serve to make voters' choices more meaningful overall, not less.

F The Information Will Be Misunderstood or Distorted

Finally, it might be objected that requiring disclosure will harm the public's interest in a rational and informed political process. Given the complexity of medical information, the public might respond irrationally to even highly relevant information, and the media or a candidate's opponents might distort such information. Our response is four-fold. First, this worry is largely mitigated by the narrowness of our standard; people simply would not have the opportunity to react irrationally to minor medical conditions. Second, just about any information about a candidate can be distorted or interpreted irrationally. A candidate's senatorial voting record can be distorted and interpreted irrationally, but democracy is clearly better served, on balance, by requiring disclosure of that information. Third, even if the public responds to information irrationally, it is still something to which members of the public have a right, and if they have a right to it, then the mere fact that they will not use it rationally does not justify withholding it. And finally, in recent instances of candidates publicizing their medical diagnoses, such as Rudy Giuliani's early-stage prostate cancer, responsible journalists have enlightened the public with well-researched news articles explaining the condition, its treatment, and prognosis. There is always the possibility that voters will misunderstand medical facts (or any facts), but the way to resolve that problem is through education, not by withholding the information in the first place.

V INSTITUTIONAL MECHANISMS

What institutional mechanism would be the best way of protecting both the democratic rights and the privacy interests is a separate, though related, question. Although a full discussion of all of the legal and institutional issues is beyond the scope of this paper, we will briefly address several possibilities here.

One possible mechanism is that before the primary elections, candidates would undergo a standardized history, physical examination, and any appropriate laboratory tests. This information along with the candidate's entire medical record would be distributed to a bipartisan panel of physicians, appointed by the president and confirmed by Congress, with members drawn from a wide variety of medical disciplines. The panel would then decide if further tests are needed, and generate a report that would list any medical conditions likely to seriously undermine the candidate's ability to perform the core duties of office, along with an explanation for why they thought the conditions listed were likely to do so.

The report would then be given to the candidate and his or her physicians to review for any errors or disagreements. The panel would take their review under advisement and make any revisions it deems appropriate. Ultimately the panel would have the final authority regarding the report's content. The candidate would then receive a final copy prior to public release. If the candidate consented to disclosure at that point, the report would be disseminated to the public. If the candidate refused to consent, then the report would be treated as part of the candidate's confidential medical record.

Disclosure to the panel could be either legally required or legally voluntary. If Congress wanted disclosure to be legally required, it might make such disclosure a filing requirement, applicable to all candidates who wish to be on the ballot. Alternatively, Congress could make such disclosure a condition of receipt of federal campaign funds. This was in fact suggested by several politicians after Tsongas's health issues become public (Altman, 1993). There are, however, constitutional limits on Congress's power to impose restrictions on candidacy, and it is an open question whether this would pass constitutional muster.

The remedy when candidates fail to disclose could simply be bad publicity and monetary fines (as it is with failing to file financial reports) (Ethics in Government Act, 1978). More strongly, intentional failure to disclose all required information to the panel, if discovered prior to the election, could be penalized by disqualification of the candidate. Once elected, though, it is unlikely that such a drastic remedy as the removal of a sitting president would be warranted, and would possibly have the same disruptive effect that a disclosure requirement seeks to avoid, viz., effectively undermining the rights of those who voted for that candidate.

It may be that disclosure would have to be voluntary. Even so, Congress could still play a role by setting standards for what counts as full disclosure, and by setting up an institutional mechanism, such as the one outlined above, for standardizing disclosure in a way that minimizes unfairness to candidates. Participation by some candidates would put public pressure on other candidates to participate as well. Note that by setting such standards Congress would avert the possibility of a disclosure one-upping, where some candidates try to force others to disclose minor health issues in order to score cheap points by being slightly more fit. ¹⁵

A voluntary standard has the advantage of providing an additional mechanism to keep potential abuse in check. Should the panel make spurious judgments, candidates would still be able to opt out of disclosing a negative report, and future candidates could boycott the procedure until a consensus had been reached on the panel's objectivity. Put another way, if participation is contingent on respect for the institution, panelists will have an additional reason for earning and maintaining that respect.

Finally, two other groups could use the standard articulated here. First, the media could use that standard to decide whether to report medical information to which they have access. Journalists can presumably disclose even if they cannot get prior consent from the candidate, but the standard gives them a guide for when doing so is justified in light of the competing rights and interests. Second, while candidates' personal physicians and their healthcare institutions would presumably need prior consent to release information, they could still use our standard to decide, in conjunction with the candidate, what should be disclosed and what should not.

VI CONCLUSION

People do not lose their right to medical privacy solely by becoming candidates for president, and their interest in that privacy is substantial despite running for public office. (Indeed, it is in many ways more substantial precisely because of their running for office.) However, candidates are in a special position in which privacy conflicts with a fundamental democratic right—the people's right to be governed only by consent, as realized in their right to vote. We have argued that the right to vote includes the right to the information necessary to make informed voting decisions, and that this right places a moral requirement upon candidates to disclose information about medical conditions that are likely to seriously undermine their ability to fulfill the core functions of the office.

ACKNOWLEDGMENTS

We would like to thank Alta Charo, Norm Fost, Pilar Ossorio, and anonymous reviewers at the Journal of Medicine and Philosophy for their helpful comments.

Notes

1. Although many of the reasons supporting disclosure in the case of presidential candidates would apply to other candidates for other offices, these are presumably of less importance. We have therefore chosen to restrict our attention to presidential candidates. But see Section IV.A, in which we briefly discuss the 25^th Amendment and presidential successors.

2. Two points are worth clarifying here. First, except where noted otherwise, by “right,” we mean “moral right.” By this we mean that to say someone has a right to something is to say something weaker than that they must, all things considered, be provided that thing. However, violating the right must be justified by more than a mere incremental increase in overall utility. Second, while the Constitution does not expressly provide citizens the legal right to vote for president, the legal right of citizens to vote for president is constitutionally protected once the state has decided to choose electoral delegates by popular vote. For a discussion see Bush v. Gore (2000, p. 98). Moreover, even if it would be constitutionally permissible for a state to decide to choose its electoral delegates by some means other than popular vote, it seems obvious that it would not be morally permissible to do so precisely because the people would still have the moral right to vote.

3. Certain problems arise from this way of formulating democratic consent. For example, suppose that a candidate wins an election on the basis of intentionally false campaign promises. While doing so does not nullify the election and the legitimacy of their holding office, it does undermine the meaningfulness of the community's consent. Similarly, consider the possibility that a candidate takes office on the basis of a miscount; the person still legitimately holds office despite the weakness of community consent. Whether or not these candidates ought to retain office will depend upon how severe the burdens of recall or impeachment would be.

4. For further discussion, see Faden and Beauchamp (1986, esp. pp. 274–297).

5. These examples raise the possibility of circumstances where a doctor would be morally required to not practice medicine and indeed should be legally prevented from doing so. By analogy, then, one might argue that there could be circumstances in which a presidential candidate suffering from a serious illness would be morally required to not seek office, and indeed, should be legally prevented from doing so. Although we agree that there could be circumstances in which a candidate is morally required not to run, we do not think that it would be warranted to legally prevent a candidate from running due to health concerns. The crucial disanalogy is that the minimal qualifications for practicing medicine are much more easily defined and less controversial than the minimal qualifications for presidential candidates. Moreover, prohibiting candidates from running would arguably diminish the rights of voters (and would likely require a constitutional amendment). This contrasts with information disclosure requirements, which augment, rather than diminish, voters' rights. See Section IV.E below. Thanks to an anonymous reviewer on this point.

6. Annas (2000), for example, appeals to “reasonable medical certainty.” Thanks to an anonymous reviewer on this point.

7. We wish to be clear that by “competent” we are referring very narrowly to the kinds of clinical judgment that doctors or psychiatrists make when they find a patient to be able to meet the minimum standards of decision-making capacity, and not competence in the sense of the ability to excel at a task. For further discussion of the concept of competence see Buchanan & Brock (1986).

8. Annas (2000) has also proposed an intermediate threshold, to which ours is similar. He says that “a candidate's medical records should be disclosed only if there is a reasonable medical certainty that a presidential candidate will not survive a four-year term, or will be unable to function mentally.” Our view differs from Annas's in several respects. First, our threshold for disclosure allows for the possibility that conditions other than four-year mortality or mental incompetence are sufficiently relevant to the candidates' ability to carry out the core functions of the office. It is thus likely to lead to broader disclosure requirements. For example, if it was known that a candidate would be physically unable to travel to other countries, this would likely be relevant in our view, but not in Annas's. Moreover, our view states a sufficient condition for disclosure requirements and includes a positive argument, in terms of democratic legitimacy and informed consent, for those requirements. Annas only provides a necessary condition and focuses on the pernicious effects of excessive disclosure.

9. The authors found no references to Cheney having disclosed his medical records during the 2004 campaign.

10. Just as the Americans With Disabilities Act (ADA) only allows employers to require medical tests when the test is relevant to “job-related functions” (ADA, 2004).

11. For further discussion, see Dworkin (1977) and Wertheimer (1983).

12. And it may well not go smoothly: even with the 25^th Amendment, a president who arrives by succession rather than by being elected president is likely to garner less public and Congressional confidence and support, and their legitimacy is likely to be contested. See Neustadt (2001) and Wicker (2001).

13. For an extended discussion of what the criteria should be for the application of the 25^th Amendment and which medical conditions meet the criteria, see Abrams (2001) Post (2001), and Drachman (2001).

14. For a useful discussion, see Dworkin (1977).

15. For further discussion of this worry, see Annas (1995).

Americans with Disabilities Act of 1990, 42 U.S.C. § 12112(d)(2)(B) (2004).

Bullock v. Carter, 405 U.S. 134 (1972).

Bush v. Gore, 531 U.S. 98 (2000).

Ethics in Government Act of 1978, 5 U.S.C. App. 4 Sec. 104(a) (2004).

Prothero, M. (2000, November 28). ‘Cheney claims good health, refuses to release medical records,’ UPI. Retrieved August, 24, 2004, from Lexis-Nexis Academic Universe database.