*280 B. General Barriers to the Effectiveness of Civil Rights Approaches to Physician Bias

1. Lack of Coherence and Comprehensiveness

The previous section's discussion of the various federal statutes that a plaintiff might employ to seek redress for a biased medical decision itself demonstrates a primary barrier to the efficacy of using civil rights statutes to address physician bias: The statutes themselves are diverse and characterized by variation with respect to the prohibited basis for discrimination, the entities subject to the law and enforcement mechanisms. Consequently, any attempted civil rights response must be tailored to the particular “flavor” of physician bias at issue in a specific case. The case of a racially biased medical decision prompts a different response from the case of gender-biased decision making, which prompts a different response from cases of disability or age-biased decisions. Thus, any response to the problem of physician bias grounded on federal civil rights laws will lack coherence. Moreover, because many instances of biased medical decisions do not constitute even an arguable violation of civil rights laws (for example, racially biased decisions made by a physician who receives no federal funding, or gender-biased decisions by most physicians), existing federal anti-discrimination laws do not offer a comprehensive approach to the problem of physician bias.

Admittedly, whether we view the civil rights approach's lack of coherence and comprehensiveness as problematic depends on our perspective on the problem of physician bias. If we view physician bias as objectionable only when the characteristic generating bias is a characteristic that also defines prohibited grounds for discrimination under federal law (e.g., race, sex, disability and age), and if we are satisfied that existing anti-discrimination laws sufficiently address those objectionable instances of biased medical decisions, only then would lack of coherence and comprehensiveness be untroubling. In other words, if we understand physician bias not as a general problem, but instead as a set of problems in specific contexts, then a set of legal responses targeted to those contexts may in fact be desirable. On the other hand, one who views biased medical decisions inclusively, as a general problem--albeit having particularly problematic manifestations for specific groups of patients--will find existing civil rights laws to be sorely inadequate for addressing the problem.

2. Challenges of Proving Intentional Discrimination

The lack of coherence and comprehensiveness of a civil rights approach to biased medical decision making calls into doubt that approach's effectiveness as a systematic response to physician bias. It is of no concern, however, to a patient whose claim of physician bias falls into one of the pockets of federal civil rights protection described above. To illustrate, an African-American Medicaid recipient who believes that his race influenced his doctor's choice for treating the patient's cardiac disease can easily *281 plead race discrimination by a recipient of federal funding and would encounter no coherence or comprehensiveness problems. This patient nonetheless faces a typically daunting--often well-nigh impossible--challenge of proving intentional discrimination on the doctor's part.

This challenge is comprised of two aspects. First, the plaintiff must prove that discrimination in fact occurred--that he received different medical treatment because of his race (or other prohibited characteristic) and not for some other reason. Second, at least under Title VI, the plaintiff must prove that the physician's discrimination was “intentional.” Although these two requirements are intimately related and typically are not assessed separately by courts, they may raise slightly different issues in the context of biased medical judgments. While courts have developed slightly varying standards for proving violations of the different anti-discrimination statutes discussed above, this section focuses on the challenge of proving a Title VI violation.

a. Proving Different Treatment Based on Race

A plaintiff who alleges that bias with respect to one of her personal characteristics affected her physician's choice of treatment is essentially claiming that her physician provided her with a treatment different from what he would have provided in the absence of that characteristic. Thus, an individual patient's claim of biased medical treatment is a claim of disparate treatment: The treatment choice was different because the patient was black, female, elderly or developmentally disabled. These claims are not likely to be framed effectively as disparate impact claims, which typically complain of the adversely disproportionate impact of a policy or practice that is facially neutral. Although some medical treatment decisions may be influenced by such policies or practices employed by institutional providers, payers or regulatory bodies, individual physicians *282 making treatment decisions typically form their decisions on a patient-by-patient basis without employing blanket policies directing the treatment choices.

How can a patient who alleges such different treatment prove that she in fact received different treatment because of her race? For example, let us hypothesize that a young black woman went to a hospital's emergency room with a broken leg. There, despite her complaints of significant pain, the emergency room doctor failed to provide her with any pain medication. Following the incident, she sues for a violation of Title VI, asserting that the physician's failure to provide her with adequate analgesia was based on her race. Courts hearing discrimination claims allow plaintiffs to prove their cases either by using direct evidence of discrimination or by producing circumstantial evidence of discrimination. Direct proof in the hypothetical case would be evidence that the emergency room physician, in deciding whether to prescribe pain medication, in fact weighed the patient's race as a factor in his decision. It is hard to imagine that a plaintiff often would be able to prove that the physician told her (or anyone else) that he was ordering or denying a particular treatment because she *283 was black. Nor is it likely that a physician's notes in a patient's medical chart would explicitly draw such a causal connection. Thus, direct evidence of different treatment based on race is likely to be quite rare.

Lacking direct evidence of race-based differential treatment, the patient may resort to circumstantial evidence to prove racially based discrimination. Circumstantial evidence allows a fact finder to infer from directly proven facts that the legally significant fact also exists. In essence, a case based on inferential evidence asks the factfinder to conclude that discrimination is the most plausible explanation for a particular proven set of facts. The Supreme Court has established a structure for litigating individual disparate treatment cases based on inference. Although the well-known McDonnell Douglas burden-shifting paradigm was developed in an employment discrimination case under Title VII of the Civil Rights Act, courts have adapted it to different contexts and different statutory schemes, including Title VI.

*284 So, to continue the hypothetical case described above, even if the patient who was denied analgesia for the pain from her broken leg cannot produce direct evidence that the doctor withheld analgesia because of her race, she nonetheless can seek to prove that her race is the most plausible explanation for the denial. She might seek to show that when other non-black patients came to the emergency room with a fractured leg and complained of pain, the defendant physician ordered analgesia. Statistical proof that this doctor tends to order pain medication only for non-black emergency room patients with fractured limbs could satisfy the plaintiff's burden of showing a prima facie case of discrimination. Similarly, she might seek to introduce evidence of statements or actions by the physician demonstrating that the physician harbored prejudice against African Americans or that he entertained stereotypes about their proclivity towards drug-seeking behavior.

Of course, under the burden-shifting paradigm the defendant physician retains the opportunity to present evidence of a non-race-based reason for failing to prescribe analgesia. For example, the physician might assert that the plaintiff did not complain of pain forcefully enough to draw his attention; he might claim that the patient was intoxicated at the time that she came to the emergency room and that therefore analgesia was contraindicated; or he may allege that the patient was uninsured and that the emergency department had a policy of not administering analgesia to uninsured patients who could not give proof of their ability to pay. Even if the physician articulates a legitimate, non-discriminatory reason for failing to prescribe analgesia, the plaintiff patient still has the opportunity to try to convince the trier of fact that the proffered reason was not the real reason, but was instead a pretext for discrimination.

*285 Describing the process by which a plaintiff might seek to prove a Title VI violation using inferential proof is straightforward enough, but such a plaintiff is likely to encounter several hurdles in trying to complete the above course. First, in the likely absence of a “smoking gun” showing direct evidence of racially motivated different treatment or evidence of the physician's generalized prejudice, a patient will need to employ statistical evidence to make her case. While statistical evidence can be difficult and expensive to develop in any case, the difficulty is multiplied when the necessary data involve the medical treatment of other patients.

As discussed above, courts are generally unwilling to allow discovery of the medical records of non-party patients in medical malpractice litigation brought by a patient. Some courts, though, will allow discovery if the plaintiff makes the case that information contained in non-party medical records is relevant and that all identifying information regarding the non-party patient can be redacted. In those cases, the court may find that the balance between the patient's need for the information and the risk of violating the confidentiality of medical records tips in the plaintiff's favor. In addition, a court may conduct the weighing process somewhat differently in a case alleging the violation of a civil rights law, for then a public interest must be factored into the calculation.

Although research reveals no published decision addressing the discoverability of patient records in a civil rights action, courts have addressed an analogous issue in the context of discrimination claims brought by physicians who have been subjected to adverse peer review action. The records of medical peer review bodies are generally protected from discovery by state law evidentiary privileges; the purpose of this protection is both to encourage physicians to participate in reviewing the competency and professionalism of their peers and to enable the free and frank sharing of information as part of that process. In a number of *286 recent cases, however, courts have found that the medical peer review privilege does not prevent discovery of peer review records in federal actions alleging civil rights violations. Indeed, in one case in which a physician alleged that the termination of his staff privileges was motivated by his race and national origin, the court allowed the physician to discover not only his own peer review records, but also records containing peer review information regarding other physicians at the hospital. In declining to recognize the evidentiary privilege in this context, the court reasoned:

The evidence [the plaintiff] seeks is crucial to his attempt to establish that he has been the subject of disparate treatment on the basis of race and ethnicity. To prove his allegations of disparate treatment, [the plaintiff] must compare the proceedings in his case to those involving similarly situated physicians. The interest in facilitating the eradication of discrimination by providing perhaps the only evidence that can establish its occurrence outweighs the interest in promoting candor in the medical peer review process. A plaintiff seeking to discover evidence of a physician's treatment of other similarly situated patients in a civil rights action thus may be able to effectively analogize his request for discovery to the peer review context in which courts have refused to protect records from discovery.

Yet the difficulty of obtaining evidence from which discrimination might be inferred is not the only barrier to recovery faced by a patient-plaintiff suing a physician for biased medical treatment under a federal anti-discrimination law. Even if the plaintiff is able to present statistical evidence derived from patient records showing, for example, the physician's pattern of treating black patients differently from non-black patients, the physician-defendant undoubtedly will assert that some factor, or combination of factors, other than race led him to choose the treatment provided to this particular patient. Just as the complexity of clinical decision making makes it difficult to eliminate potentially confounding variables *287 in researching the causes of health treatment disparities, so too does that complexity complicate the civil rights plaintiff's task in isolating race (or any other group characteristic) as a causative factor. Because medicine is an inexact science and physicians typically take a wide variety of facts into account in exercising clinical judgment, the physician's assertion of a legitimate non-discriminatory reason for the selected treatment will likely prove difficult to rebut without some direct evidence of bias. Thus, even a plaintiff who is allowed to admit all available evidence suggesting discriminatory medical treatment often will fail to persuade the fact finder that any treatment differential was in fact based on race, rather than some other factor.

b. Proving that Discrimination was Intentional

The foregoing discussion has focused on the challenge inherent in trying to prove that a patient's race did in fact influence a physician's decision regarding the medical treatment provided to (or denied) the plaintiff. But even a patient-plaintiff who can convince a court that her physician provided her with different medical treatment because of her race will not necessarily succeed on her civil rights claim. In order to succeed, she must also prove that the discrimination was intentional. The Supreme Court recently reaffirmed that a private plaintiff suing to enforce Title VI must show intentional discrimination. According to the Court, Title VI itself “proscribes only those racial classifications that would violate the Equal Protection Clause or the Fifth Amendment.” Consequently, a plaintiff must point to evidence of racially discriminatory intent or purpose. So how can a plaintiff show that her physician, when ordering different medical treatment for the plaintiff because of her race, intentionally discriminated?

Admittedly, the question posed rings confusing: After all, isn't all disparate treatment racial discrimination intentional? How can one person unintentionally treat another differently because of the other's race? Doesn't the causal relationship indicated by “because of” also demonstrate *288 the actor's intent to discriminate? In many cases, the answer to the final question is “yes.” In others words, the actor found to have considered race in making a decision is also found to have done so consciously and, therefore, intentionally. If she knows she is taking race into account in making her decision, then she is intentionally discriminating. But what if she is not aware that she is taking race into account?

The question of what exactly the Supreme Court means when it requires proof of “intentional” discrimination has occupied a number of scholars and produced a rich literature examining whether liability for violating anti-discrimination laws should extend to actors whose acts of differential treatment flow from subconscious biases. The seminal work in this area is Charles Lawrence's The Id, the Ego, and Equal Protection: Reckoning with Unconscious Racism, which draws on psychiatric and psychological literature to argue that most contemporary racism is embodied more in unconscious stereotyping than in conscious bigotry. Lawrence recognizes that existing jurisprudence effectively relegates acts influenced by unconscious racial stereotyping to a sort of intent limbo:

Traditional notions of intent do not reflect the fact that decisions about racial matters are influenced in large part by factors that can be characterized as neither intentional--in the sense that certain outcomes are self-consciously sought--nor unintentional--in the sense that the outcomes are random, fortuitous, and uninfluenced by the decisionmaker's beliefs, desires, and wishes. Other scholars have built on Lawrence's work to examine the legal treatment of discrimination that might be deemed “negligent” or “accidental.”

A central premise of much of this literature reflects theories, developed by cognitive psychology, that characterize stereotyping as a natural, and indeed predictable, human response to the richness of information and stimuli in the surrounding environment. Thus, according to cognitive *289 theory, people whose mindset includes stereotypes or biases are not necessarily “prejudiced” against those groups about whom their minds have formed biases; indeed, people are unlikely to be aware of the bias's existence or its influence on their decisions. While a recent review of social science research into stereotyping and discrimination indicates that more recent studies find that motivation and affect may also contribute to unconscious discrimination, the reviewer still concludes: “No matter how the interaction of cognitive, affective and motivational factors occurs, social science research demonstrates that race and gender bias and prejudice resulting in discriminatory behavior are the result of unconscious, as well as conscious, phenomena.” According to these scholars, much of contemporary racial discrimination occurs without the discriminating decision maker being aware of race's influence on his decisions.

Most of the legal scholars who have concentrated their attention on the legal status of unconscious discrimination have done so in the context of asking whether liability for employment discrimination under Title VII should extend to unconscious disparate treatment of employees on statutorily forbidden grounds. These scholars have reached varying conclusions on that question, and the courts have not clearly resolved the question either. Moreover, although Charles Lawrence's article itself *290 focused on subconscious racism's role in equal protection analysis, the issue of whether unconscious discrimination might qualify as intentional discrimination sufficient to show a violation of Title VI and other Spending Clause legislation has not been extensively studied.

A full exploration of whether courts should find a Title VI violation when a recipient of federal funding treats a person differently because of his race, but is not conscious of the racial motivation, lies beyond the scope of this Article. This is the very question, however, that a court hearing a Title VI claim alleging racially biased medical decision making may be called upon to answer. It is doubtless possible that some physicians making clinical decisions consciously take a patient's race (or sex, age or disability) into account in a manner irrelevant to the patient's need for medical treatment. These physicians would appear to discriminate “intentionally” under any meaning of the word. Most commentators in the medical literature, however, concur that--to the extent that physician bias plays a role in producing documented disparities in the levels of different treatments provided--the bias is probably largely subconscious, so that physicians' judgments may be influenced by biases without the physicians' being aware of that influence.

*291 What result then follows--to return to the hypothetical posed earlier--if the fact finder concludes from the evidence that no factor other than race adequately explains why the emergency room physician failed to provide adequate analgesia to the black woman with a broken leg, but the fact finder also believes that the physician may not have been consciously aware that the patient's race influenced his decision? Scholars have laid the theoretical foundation for arguing that motivation or causation should be the determinative factor in identifying “intentional discrimination” (i.e., that discrimination is intentional if race motivated or played a causal role in producing the complained-of behavior) and that a plaintiff alleging discrimination should not be required to prove a conscious decision to discriminate. Whether a plaintiff alleging biased medical treatment could succeed in persuading a court to construct a decision upon this foundation, however, is another matter. Thus, even a plaintiff who can garner evidence sufficient to convince a court that her race influenced her doctor's choice of medical treatment may not be able to persuade the court that her physician's discrimination was intentional.

3. Judicial Reluctance to Intrude on Doctors' Judgment and State Regulation

A final problem with any attempt to invoke federal civil rights laws in response to biased medical decisions lies in the courts' reluctance to apply federal anti-discrimination law to a claim that the court views as essentially alleging poor quality or inappropriate medical care. This reluctance flows from two related sources: First, the courts are disinclined to take on the task of second-guessing professional medical judgment. Second, to the extent that a claim entails assessing the quality of care provided, courts emphatically proclaim that any such assessment should be conducted not as a matter of federal law, but as a matter of state law regulating the quality of medical care, i.e., as a medical malpractice claim.

As an initial matter, one might wonder why courts would characterize an action alleging biased medical decisions as essentially a complaint about poor quality care. This characterization, however, is in some sense accurate. It stands to reason that a patient who believes that she received superior medical treatment because of her race (i.e., she was discriminated “in favor of,” rather than “against”) is unlikely to sue her physician. *292 Accordingly, complaints of discriminatory medical judgments are likely to include--as part of the claim that the plaintiff was treated differently from other similarly situated patients--a claim that the treatment received was somehow inferior to and of poorer quality than that received by other patients. Indeed, commentators too have characterized evidence of racial disparities in medical treatment as evidence of poor quality of care. Recognition of the quality aspect implicit in a claim of discriminatory medical treatment, however, does not necessarily compel a court's unwillingness to address the alleged discrimination. So what explains the courts' reluctance?

One explanation lies in the courts' discomfort with questioning the soundness of medical judgment reflected in a medical treatment decision under any rubric other than medical malpractice. This discomfort is evident in cases alleging that medical choices violated disability discrimination laws. For example, in deciding that Section 504 of the Rehabilitation Act did not extend to medical decisions regarding disabled newborns, the Second Circuit rejected the government's assertion that a disabled newborn who had been denied life-saving surgery had thereby been subjected to discrimination. The court found that characterizing the medical decision as discriminatory tookan oversimplified view of the medical decisionmaking process. Where the handicapping condition is related to the condition(s) to be treated, it will rarely, if ever, be possible to say with certainty that a particular decision was ‘discriminatory’ . . . . Beyond the fact that no two cases are likely to be the same it would invariably require lengthy litigation primarily involving conflicting expert testimony to determine whether a decision to treat, or not to treat . . . was based on a ‘bona fide medical judgment’, however that phrase might be defined. Some other courts have followed the Second Circuit's lead and have simply declined to apply disability discrimination law to medical treatment decisions.

*293 Moreover, even those courts willing to entertain claims that medical choices violate disability discrimination law have shown significant deference to physicians' exercise of medical judgment. For example, in the recent case Lesley v. Hee Man Chie, the First Circuit addressed directly the question of the appropriate level of respect to be paid to medical decisions and sought to steer a middle course between “a rule giving physicians complete deference and a rule requiring a full-fledged inquiry into their diligence.” The “middle ground” standard chosen, however, remains quite deferential, allowing the plaintiff to challenge a medical decision under the Rehabilitation Act only if she can “show[ ] the decision to be devoid of any reasonable medical support.” As a practical matter, because of the wide variations in medical treatment patterns, only the most egregiously discriminatory (or negligent) medical decision is likely to pass the test of being “devoid of any reasonable medical support.” Under this standard, if a physician can point to any reasonable medical support for her treatment choice, she avoids liability even if bias in fact influenced her choice.

The court's analysis in Lesley also highlights the second aspect of courts' lack of enthusiasm for applying federal anti-discrimination laws to medical treatment decisions. Not only do the courts wish to avoid second-guessing a medical judgment, they also want to avoid applying federal statutes to an area of traditional state regulation: medical malpractice law. As partial justification for its refusal to scrutinize closely the defendant doctor's medical decision, the Lesley court reasoned:

[C]ourts should not probe so far into a doctor's referral decision as to inquire whether it was the correct or best decision under the circumstances, or even whether it met the standard of care for the profession. Lest questions of medical propriety be conflated with questions of disability discrimination, it must take *294 more than a mere negligent referral to constitute a Rehabilitation Act violation. Thus, if a plaintiff alleging a discriminatory medical treatment decision stresses too strongly the inferiority of the treatment provided, he risks the court's rejecting his claim as an attempt to federalize malpractice law.

This judicial discomfort with applying federal statutes to patients' claims that include any complaint about the quality of medical care received is not confined to the realm of anti-discrimination claims. For example, courts hearing cases alleging violations of the Emergency Medical Treatment and Active Labor Act --the federal “anti-dumping” law for hospitals--have consistently proclaimed that the Act was not intended to create a federal malpractice action for negligent screening or treatment in hospitals and thus does not apply to complaints of negligent or poor quality care provided by emergency department personnel. Similarly, the Supreme Court, in rejecting a patient's claim that a managed care organization's use of financial incentives for physicians breached fiduciary duties under the Employee Retirement Income Security Act, recently emphasized the risk of replicating state malpractice actions: “[F]or all practical purposes, every [such] claim of fiduciary breach by an HMO physician . . . would boil down to a malpractice claim, and the fiduciary standard would be nothing but the malpractice standard traditionally applied in actions against physicians.” In light of these cases, it would hardly be surprising to find courts presented with federal civil rights actions alleging discriminatory *295 medical decisions concluding that the plaintiff's real complaint is with the quality of medical care received and that such complaints are sound in state medical malpractice law and thus are not properly adjudicated under federal statutes.

Both facets of the courts' reluctance to apply federal civil rights laws to medical decisions also appear in the passive stance that federal regulators have adopted towards applying those laws to treatment decisions. The U.S. Commission on Civil Rights, as part of its 1999 report on health care disparities, reported on the policy development and rulemaking processes of DHHS's Office of Civil Rights (OCR) with respect to Title VI enforcement. One failing that the Commission noted was OCR's failure to articulate clearly what “discrimination” in the health care context means, particularly when it comes to determining what “equal access to quality health care” requires. From the Commission's viewpoint, equal access requires that all Americans receive medical care of equal quality. However, OCR's legal staff questioned both the agency's jurisdiction over denial of equal access claims and the agency's ability to assess the equality of access so defined, citing an inability to judge what appropriate care is. As the Commission reported: “Several staff noted the inability to obtain the medical expertise needed to challenge a [health care provider's] medically related rationale, suggesting the OCR was not equipped to make such determinations.” Thus, the government agents charged with enforcing the civil rights laws in health care settings appear to share the courts' reluctance to get involved in cases alleging discrimination in medical treatment.

This reluctance may pose an insurmountable obstacle for a civil rights plaintiff alleging discriminatory medical treatment. A court that detects in her complaint (or her proof) an allegation that the plaintiff received inferior medical care may well send her packing to state court with instructions to file a malpractice claim, notwithstanding the plaintiff's central allegations of discriminatory treatment. Because of the improbability, discussed in Part IV, that medical malpractice law will provide many such *296 plaintiffs with an avenue for redressing biased decisions, however, these instructions may leave a patient whose physician made biased treatment decisions without any legal recourse.