Indiana Law Review


Indiana Law Review
Volume 31 1998 Number 1

ARTICLES

End-of-Life Decision Making: What We
Don't Know, We Make Up; What We Do

Know, We Ignore

Sandra H. Johnson*

Introduction

Imagine that well-structured empirical studies consistently indicated that

doctors do not tell patients what tests they are performing or why;' imagine that

doctors can frame the information they provide patients and quite successfully

generate the physician-desired consent or refusal of the treatment;^ and imagine

that only about half of patients recall being informed of serious risks of

interventions, such as the risk of death.^ Now, imagine one more thing: that
outcome studies indicate that survival rates do not vary according to whether the

physician informed the patient of significant risks, benefits and alternatives of

treatment."*

* Professor of Law, St. Louis University School of Law. A.B., St. Louis University; J.D.,

New York University; LL.M., Yale Law School.
1. Lori B. Andrews, Past as Prologue: Sobering Thoughts on Genetic Enthusiasm, 27

Seton Hall L. Rev. 893, 904 (1997) (describing practice of clinical genetic testing without

disclosure to patients); Thomas Koenig & Michael Rustad, His and Her Tort Reform: Gender
Injustice in Disguise, 70 WASH. L. REV. 1, 69-70 (1995) (detailing cases in which physician failed

to inform female patients that certain surgical procedures would be done).

2. See, e.g., Theresa M. Marteau, Framing ofInformation: Its Influence Upon Decisions

of Doctors and Patients, 28 BRIT. J. Soc. PSYCHOL. 89, 93 (1989) (stating that options framed

positively were more likely to be chosen than similar options framed negatively); David Orentlicher,

The Illusion of Patient Choice in End-of-Life Decisions, 267 JAMA 2101 (1992) (stating that
patient preferences regarding treatment options are often shaped by their physician's values); Cathy

J. Jones, Autonomy and Informed Consent in Medical Decisionmaking: Toward a New Self
Fulfilling Prophecy, 47 WASH. & LEE L. REV. 379 (1990) (describing bias in information provided).

3. See, e.g., Terence C. Wade, Patients May Not Recall Disclosure of Risk of Death:

Implicationsfor Informed Consent, 30 MED. SCL L. 259, 260 (1990) (discussing a study of patients'

failure to recall disclosure ofthe risk of death). But see Alan Meisel & Mark Kuczewski, Legal and
Ethical Myths About Informed Consent, 1 56 ARCHIVES INTERNAL MED. 252 1 , 2524 ( 1 996) (arguing

that recall is not the equivalent of understanding and that understanding at the point of decision

making is the critical issue).

4. Informed consent has not been proven to affect survival rates. However, it has been



14 INDIANA LAW REVIEW [Vol. 31:13

If good empirical data revealed all of this about actual practice, what
response would be most appropriate? Would we abandon the ethical and legal
duty of informed consent as impractical and unworkable and require only that the

doctor make treatment choices in the patient's best interest? Should we?
A serious challenge to the law and public policy supporting a requirement of

informed consent has not yet materialized,^ even though all but one of the data

statements described above are authentic. Repeatedly, empirical studies have

cast doubt on the practicality and performance of informed consent,^ but the

place of that value in law and ethics has not yet collapsed, and may never
collapse, in response to the data.^

To what extent should data on actual practices determine public policy and
legal and ethical standards in health care decision making? A first-level response
to this question is whether the particular study has produced accurate and

verifiable results and, assuming that the research was itself well-designed and
that the data are solid, whether the results are being used within the boundaries

and limitations of the study's design. All statistical evidence generates these

questions. In fact, the conference at which the papers in this Symposium issue
of the Indiana Law Review were originally presented was entitled Lies, Damn
Lies and Statistics?: How Empirical Research Shapes Health Law and Policy}
The conference title quoted Mark Twain's identification of "three kinds of lies;"^
but, in addition to providing a catchy title for a conference,'^ Twain's cutting

associated with otlier effects. See, e.g., Wendy Levinson et al., Physician-Patient Communication:

The Relationship With Malpractice Claims Among Primary Care Physicians and Surgeons, 277

JAMA 553 (1997) (discussing patient dissatisfaction and litigation resulting from breakdowns in
physician-patient communications). In a substantial analysis of informed consent, Peter Schuck

observed that studies on the costs of informed consent are lacking, but that there are reports of other

positive outcomes including: achieving better results, decreased likelihood of malpractice claims,

and enhanced perception of competence and control by the patient. Peter Schuck, Rethinking

Informed Consent, 1 03 YALE L.J. 899, 943 ( 1 994).

5. Schuck, supra note 4, at 902-03.

6. Meisel & Kuczewski, supra note 3, at 2521-22.
7. Schuck, supra note 4, at 902-03. In reference to this debate, Schuck identified three

different versions of informed consent doctrine: the "letter and spirit of the doctrine" ("law in

books"); the law as imagined or caricatured by some doctors ("law in the mind"); and the doctrine

as actually practiced ("law in action"). Id. at 903. Schuck noted that commentators fall into two

groups—^the idealists and the realists—and that these groups talk past each other in discussions of
informed consent. Id. at 903-04.

8. Symposium, Lies, Damn Lies and Statistics?: How Empirical Research Shapes Health
Law and Policy, 31 IND. L. REV. 9 (1998).

9. Mark Twain, Mark Twain's Own Autobiography 1 85 (Univ. of Wis. Press 1 990).
In his autobiography, Twain quoted English politician and historian Benjamin Disraeli, who stated

that "there [are] three kinds of lies: 'lies, damned lies, and statistics.'" Momingstar, Inc. v. Pilgrim

Group, Inc., 29 Cal. Rptr. 2d 547, 553 (Cal. Ct. App. 1994).

10. Twain's original statement may have been closed with an ironic exclamation point, of

course; surely not by a question mark. The addition of the exclamation is a slight offense, however.



1 998] END-OF-LIFE DECISION MAKING 1

5

critique of quantitative empirical research, at a minimum, identifies the
skepticism and skill with which empirical research in bioethics should be

analyzed.

This data-checking inquiry is not in itself an adequate response especially for

any enterprise that sets standards of conduct, which includes both law and

bioethics. Even well-designed statistically valid data requires interpretation.
When verifiable data on current practices clash with well-established

normative standards, empirical research commonly appears to be appreciated
only narrowly. The data may be construed, for example, to reveal only
implementation or enforcement problems, while marginalizing the empirical

challenge and leaving the core values and norms intact. Thus, it stimulates a call

for more education, or greater "commitment," or stronger enforcement, rather
than altering the normative standards.

In rare instances, empirical research may trigger radical calls for
overthrowing the basic principles and paradigms that underlie established ethical

and legal duties. One of the most notable occasions which prompted such a
challenge occurred following investigation into actual practices in human
experimentation, exposing evidence that medical researchers in the U.S.

performed experiments on human beings in secret and without disclosure to their
subjects.'' In such a case, empirical research on actual behavior is seen as

revealing an inadequacy in basic norms.'^

As an applied ethics, bioethics must struggle with context. Empirical
research tests whether ethical and legal standards "fif the health care setting.'^

Bioethics, however, shares with law a fundamental tension in the essential norm-

setting function between maintaining desired norms despite sometimes frequent

violation and assuring that standards are realistic and rest on a substantial

practice of voluntary compliance.

The controversy over the appropriate role of empirical research in law has
persisted for over a century. Coincidentally, the Center's conference on

and the prominence accorded this commentary attests to the integrity of the groundbreaking

empirical research conducted by the Indiana University School of Law—Indianapolis Center for
Law and Health.

1 1

.

David J. Rothman, Strangers at the Bedside: A History of How Law and
Bioethics Transformed Medical Decision Making 3 (1991).

1 2. In health law, these occasions arise in the unusual case in which the courts depart from

the standard of the profession and establish a judicial standard of professional duty. The classic

case is Helling v. Carey, 519 P.2d 981 (Wash. 1974). In informed consent doctrine, at least half

of the states hold the physician to the standard of the profession rather than to the patient's

expectations. Barry R. Furrow et al., Health Law § 6- 10(a) (1995). In cases involving the
withdrawal of life-sustaining treatment, many of the earliest cases attempted to harmonize their

stated legal rule allowing termination of treatment with the ethical standards of medical practice

even though the practice of withdrawal was still controversial. See, e.g.. In re Quinlan, 355 A.2d

647, 664-69 (N.J. 1976).

13. Susan M. Wolf, Shifting Paradigms in Bioethics and Health Law: The Rise ofa New

Pragmatism, 20 Am. J.L. & Med. 395, 403-04, 409-10 (1994).



1

6

INDIANA LAW REVIEW [Vol. 31:13

empirical research took place shortly after the centennial celebrations of the

publication of Oliver Wendell Holmes' essay, The Path of the Law, in the
Harvard Law Review}^ In his essay, Holmes claimed: "For the rational study
of the law, the black-letter man may be the man ofthe present, but the man of the
future is the man of statistics . . . ."^^ Holmes was not merely predicting the
influence of statistics, ^^ which had Twain lamenting only a few decades later, he
was speaking prescriptively. Holmes argued that the law should be designed so
that its actual effects take precedence over arguments based solely on morality

or history or philosophy.'^

Holmes used the criminal law to illustrate his point that the impact of the law
is the true test of the law,'^ but he might have used "law and bioethics" had it

existed a century ago. Both fields, even more so than many other fields of law,
are often viewed as embodying society's moral character,'^ or at least society's

declaration ofthe correct ordering of civilized communities.^° Both criminal law

and bioethics bear the mark of notions ofthe good and the natural order ofthings.
Holmes said the following about crime and punishment:

What have we better than a blind guess to show that the criminal law in
its present form does more good than harm? . . . Does punishment deter?

1 4. Oliver Wendell Holmes, The Path ofthe Law, 1 Harv. L. Rev. 457(1 897), reprinted

m 110 Harv. L. Rev. 991 (1997).
15. Mat 1001.
16. Holmes' insight has proven largely true in the influence of several significant

movements, including Legal Realism, Law and Society, Law and Economics, and Critical Legal

Studies, which all emerge from an empirically-based analysis of the law. Each of these schools

challenge what law is and what matters in law. Martha Minow, The Path as Prologue, 1 10 Harv.

L.Rev. 1023, 1023-25(1997).

17. Robert W. Gordon, The Path ofthe Lawyer, 1 10 Harv. L. Rev. 1013, 1013-14 (1997).

18. Holmes, supra note 14, at 1001-02.

19. See 136 CONG. Rec. 12,251, at 12,260 (1990) (statement of Sen. Grassley) ("In a

country that cherishes a separation between the state and any officially sanctioned religious practice,

the criminal law is one of the few available institutions through which society can make a moral

statement . . . ."); Sissela Bok, At the Juncture ofTheory and Practice: Remarks on Receiving the

Henry Knowles Beecher Award, 26 HASTINGS Ctr. Rep. 5 (May-June 1996); Thomas P. Griesa,

There Is No Casefor Legalizing Drugs, WALL ST. J., Aug 10, 1993, at A14 ("We should bear in
mind that the foundational fact about the criminal law is that it is a moral judgment."); George F.

Will, The Sting ofShame, WASH. POST, Feb. 1, 1996, at A21 ("[T]he criminal law's expressive

function is to articulate society's moral condemnation.").

20. See 136 CONG. Rec. 1 1,224, at 1 1,250 (1990) (statement of Sen. Hatch) ("Through the

imposition ofjust punishment, civilized society expresses its outrage and sense of revulsion toward

those who, by contravening its laws, have not only inflicted injury upon discrete individuals, but

also weakened the bonds that hold communities together.") (quoting Stephen J. Markman & Paul
G. Cassell, Comment, Protecting the Innocent: A Response to the Bedau-Radclet Study, 4 STAN.

L. Rev. 121, 157 (1988)); Edmund D. Pellegrino, The Metamorphosis ofMedical Ethics: A 30-

Year Retrospective, 269 JAMA 1 158 (1993).



1 998] END-OF-LIFE DECISION MAKING 1

7

Do we deal with criminals on proper principles? ... If the typical
criminal is a degenerate, ... it is idle to talk of deterring him by the
classical method of imprisonment. ... If, on the other hand, crime, like
normal human conduct, is mainly a matter of imitation, punishment
fairly may be expected to help to keep it out of fashion.^'

The calculation advocated by Holmes' century-old essay does not yet, and
probably never will, determine the outcome of public controversies over criminal

law. Arguments over whether the lack of deterrent effect is relevant to the use
ofthe death penalty^^ and the effect of "three-strikes" sentencing^^ evidence the

continuing struggle over the purpose of criminal law and punishment.

Continuing disputes over the causes and prevention of crime also illustrate the

intractability of answering Holmes' apparently empirical social inquiries.^"^

Current debates in the law of bioethics have a similarly complex relationship
with facts and values; rationality and irrationality; moral aspiration and

consequentialist accounting; intuition and evidence.

The debate over the relative influence of aspiration and practice, though an
old debate both in law and ethics, is being pressed now in bioethics because of
the recent emergence of a body of substantial empirical research on the operation

of bioethics in practice. The most extensive empirical examination of bioethics

"at the bedside" to date is also the most recent. From 1989 to 1994, the Robert
Wood Johnson Foundation funded an ambitious study of care for patients

21. Holmes, supra note 14, at 1002.

22. See generally MICHAEL L. RaDELET& MARGARET VANDIVER, CAPITAL PUNISHMENT IN
America (1988); Walter Berns et a!., The Death Penalty: A Philosophical and Theological

Perspective, 30 J. MARSHALL L. Rev. 463 (1997); William J. Bowers et al., A New Look at Public

Opinion on Capital Punishment: What Citizens and Legislators Prefer, 22 Am. J. Crim. L. 77

(1994); Stephen B. Bright, The Death Penalty as the Answer to Crime: Costly, Counterproductive

and Corrupting, 36 SANTA CLARA L. REV. 1069 (1996); Walter L. Gordon, III, Death Penalty:

National Disaster Visits California, 33 SANTA CLARA L. Rev. 283 (1993); Jordan Steiker, Essay,

The Long Road Up From Barbarism: Thurgood Marshall and the Death Penalty, 7 1 Tex. L. Rev.

1131 (1993).

23. See generally Meredith McClain, Note, "Three Strikes and You 're Out ": The Solution

to the Repeat Offender Problem?, 20 Seton Hall Legis. J. 97 (1996); Mark W. Owens, Legislative

Note, California 's Three Strikes Law: Desperate Times Require Desperate Measures—But Will
It Work?, 26 Pac. L.J. 881 (1995); Ilene M. Shinbein, Note, "Three-Strikes and You 're Out": A

Good Political Slogan to Reduce Crime, But a Failure in Its Application, 22 NEW Eng. J. ON Crim.
& Civ. Confinement 175 (1996); Victor S. Sze, Comment, A Tale of Three Strikes: Slogan
Triumphs Over Substance as Our Bumper-Sticker Mentality Comes Home to Roost, 28 LOY. L.A.

L.Rev. 1047(1995).

24. See generally Stephen Kinsella, A Libertarian Theory ofPunishment and Rights, 30

LOY. L.A. L. Rev. 607 (1997); Richard Lowell Nygaard, Essay, On the Philosophy ofSentencing:

Or, Why Punish?, 5 WiDENER J. PUB. L. 237 (1996); Thomas J. Philipson & Richard A. Posner, The
Economic Epidemiology ofCrime, 39 J.L. & ECON. 405 (1996); Negley K. Teeters, Fundamentals
ofCrime Prevention, FED. PROBATION, Sept. 1995, at 63.



1

8

INDIANA LAW REVIEW [Vol. 31:13

hospitalized with life-threatening conditions?^ SUPPORT, which is discussed
later in this Article, tested the actual operation of informed consent, patient

autonomy, and compliance of physicians with patient or surrogate choice in

medical treatment decision making, among other events in the treatment of these
patients.^^

SUPPORT contrsists with another type of influential study of bioethics, one
with a longer history. In 1978, Congress established the President's Commission
for the Study of Ethical Problems in Medicine and Biomedical and Behavioral

Research.^^ The President's Commission studied bioethics from the position of
logic and values and clinical experience, though largely intuitive or anecdotal

experience and rarely empirical .^^ The President's Commission produced a
document that provided the foundation for the development of the law and
principles of bioethics regarding life-sustaining treatment decisions.^^ This

document. Deciding to Forego Life-Sustaining Treatment^^ has been cited as
persuasive in no fewer than thirty-five appellate judicial opinions resolving end-

of-life treatment issues.^' In the nearly twenty years that passed between the

President's Commission and SUPPORT, the basic principles of bioethics,
especially as they were captured in law, remained grounded primarily in the

"head work" of the Commission and its progeny.^^

Although empirical research is a relative latecomer to bioethics, data on

practices relating to medical ethics have already produced sometimes troubling

questions concerning the gap between normative principles, including those

adopted by the President's Commission, and reality. For example, are the basic

values of American bioethics and the legal framework of health care decision

making culture-bound and culturally exclusive of large numbers of U.S.
patients?^^ Are surrogate decision makers, including family members, "reliable"

25. The SUPPORT Principal Investigators, A Controlled Trial to Impede Carefor Seriously
III Hospitalized Patients: The Study to Understand Prognoses and Preferencesfor Outcomes and

Risks of Treatment (SUPPORT), 274 JAMA 1591 (1995) [hereinafter ^t/P/'O/^T].
26. Id

27. 42 U.S.C. §300(1978).

28. President's Commission For the Study of Ethical Problems in Medicine and

Biomedical and Behavioral Research, Deciding to Forego Life-Sustaining Treatment

(Comm. Print 1983).

29. Id

30. Id

3 1

.

See, e.g., Thor v. Superior Court, 855 P.2d 375, 381 (Cal. 1993); Care & Protection of
Beth, 587 N.E.2d 1377, 1380 (Mass. 1992); In re Moorehouse, 593 A.2d 1256, 1261 (N.J. Super.

Ct. App. Div. 1991); In re Fiori, 673 A.2d 905, 912 (Pa. 1996); In re Guardianship of L.W., 482

N.W.2d 60, 70-73 (Wis. 1992).

32. Similar efforts to guide the development of standards and practices in bioethics include

much of the work of the Hastings Center, a private bioethics organization that has produced the

influential Guidelines on the Termination of Life-Sustaining Treatment and the Care of the Dying

(1987), among other documents, and The New York State Task Force on Life and Law.
33. See, e.g., Leslie Blackball et al.. Ethnicity and Attitudes Toward Patient Autonomy, 274



1998] END-OF-LIFE DECISION MAKING 19

in their understanding of the principal's preferences concerning medical

treatment?^"* Even if a patient executes an advance directive, is the patient's
doctor likely to comply with the document in directing medical treatment
decisions?^^ Each ofthese questions has been studied empirically, and it is from

such studies that an increasingly more robust body of empirical research in
bioethics has formed.

There has not been a clear transition of authority in bioethics between the

approach exemplified by the President's Commission of 1978 and the inquiry
represented by the 1995 report of the SUPPORT results, however. Each form of
bioethics "study," philosophical and empirical, continues to influence the area.

Neither has achieved a commanding presence. Even increased and more
powerful empirical research will not overtake the philosophical arm of bioethics,

however, because the tension between the "is" and the "ought," between the real

and the ideal, is inherent in any normative work that seeks to carry an authority
in practice.

This Article examines how the debate over the appropriate source of legal
and ethical norms in medicine has been played out publicly in recent judicial

decisions regarding the legalization of physician-assisted suicide. It begins with

an analysis of the Ninth Circuit'3 majority opinion in Compassion in Dying v.

Washington?^ It goes on to examine the results ofSUPPORT and responses to
those results. The Ninth Circuit opinion illustrates the use of empirical research

and the framing of core questions in factual terms in efforts to change the law of

bioethics. Responses to SUPPORT iWustmte the persistence of basic principles
in the face of substantially challenging empirical research.

I. COMPASSION IN Dying V. Washington

The Ninth and Second Circuit Courts of Appeals issued startling decisions
in 1996. The courts' decisions in Compassion in Dying v. Washington^^ and in
Quill V. Vacco^^ were startling simply because two federal appellate courts

recognized a new constitutional right, an exceptional result in itself in light ofthe
constitutional jurisprudence of this decade. The United States Supreme Court
granted certiorari in both of those cases^^ and heard oral arguments in January

JAMA 820 (1995); Sheila T. Murphy et al., Ethnicity and Advance Care Directives, 24 J.L. MED.
& ETHICS 108 (1996).

34. See, e.g., P.M. Layde et al., Surrogates ' Predictions of Seriously Hi Patients

'

Resuscitation Preferences, 4 ARCHIVES FAMILY MED. 518 (1995).

35. See, e.g., Marion Danis et al., A Prospective Study ofAdvance Directives for Life-

Sustaining Care, 324 NEW Eng. J. MED. 882 (1991).
36. 79 F.3d 790 (9th Cir. 1996), rev 'd sub nom. Washington v. Glucksberg, 1 17 S. Ct. 2258

(1997).

37. Id.

38. 80 F.3d 716 (2d Cir. 1996), rev'd, 1 17 S. Ct. 2293 (1997).

39. Compassion in Dying, 79 F.3d at 790, cert, granted sub nom. Washington v.

Glucksberg, 1 17 S. Ct. 37 (1996); Quill, 80 F.3d at 716, cert, granted, 1 17 S. Ct. 36 (1996).



20 INDIANA LAW REVIEW [Vol. 31:13

1997. The course of the oral arguments foreshadowed the outcome of the cases
and, as was thereafter generally expected, the Supreme Court reversed both
circuit court decisions.''^ There was no dissent, but five justices joined in
concurring opinions.'*'

The drama of a U.S. Supreme Court ruling on the legal status of assisted
suicide concluded somewhat anti-climactically. The Court neither recognized a
new or expanded right to die nor declared that the U.S. Constitution prohibited
the states from recognizing or creating such a right within their own authority."*^
Justice Stevens' concurring opinion indicated he does not believe that the Court's

decision closes the door on later consideration of a federal constitutional right to

physician-assisted suicide under different circumstances."*^ Despite Stevens'

opinion, the focus now moves to the states. State courts,"*"* state legislatures,"*^
and popular referenda/initiatives"*^ are the arenas for the next round in the

controversy over legalization of assisted suicide. As the arena shifts, the major
points of argument reflected in these judicial opinions are likely to be repeated.

This Article focuses on the Ninth Circuit's majority opinion and only less so

on the Supreme Court's majority opinion, although a note on the contrast
between the two is appropriate. Although the Supreme Court authoritatively
reversed the Ninth Circuit, the Supreme Court carried an easier burden than did
the Ninth Circuit and than will the state courts and state legislatures. The
Supreme Court did not take on the substantial burden of establishing a legal right

in the face of great controversy. Nor did it have to justify institutionalizing the
status quo under significant challenge. The Court's decision intentionally left the

ultimate question of legalization open and would not close the door on the

40. Glucksberg, 1 17 S. Ct. at 2258; Quill, 1 17 S. Ct. at 2293.

41

.

See Glucksberg, 1 17 S. Ct. at 2275 (Souter, J., concurring in judgments); Washington

V. Glucksberg, 117 S. Ct. 2302, 2303 (1997) (O'Connor, J., concurring in judgments); Id. at 2304

(Stevens, J., concurring in judgments); Id. at 2310 (Ginsburg, J., concurring in judgments); Id.

(Breyer, J., concurring in judgments).

42. Glucksberg, 1 17 S. Ct. at 2258.

43. Justice Stevens specifically stated that "there are situations in which an interest in

hastening death is legitimate .... I am also convinced that there are times when it is entitled to

constitutional protection." Glucksberg, 1 17 S. Ct. at 2305. He remained open to the possibility that

"an individual plaintiff seeking to hasten her death, or a doctor whose assistance was sought, could

prevail in a more particularized challenge." Id. at 2309.

44. See, e.g., Krischer v. Mclver, 697 So. 2d 97 (Fla. 1997) (After the Supreme Court's

decision, the Florida Supreme Court held that the Florida Constitution did not include a right to

physician assistance in death.).

45. See, e.g., S. 653, 89th Leg., Reg. Sess. (Mich. 1997); Henriette Campagne, Lawmakers

Take Up Assisted Suicide Issue, MASS. Law. Wkly., June 30, 1997, at 27.

46. Oregon is the only state that has approved the legalization of physician-assisted suicide,

but the referendum vote is currently being challenged in the Oregon state courts. Or. Rev. Stat.

§§ 127.800-.897 (1996). The "Death with Dignity Act" will be sent back to the Oregon voters in

November 1997. Richard Carelli, Court Unanimously Rules Against Doctor-Assisted Suicides,

Det. Free Press (June 27, 1 997) <http://www.freep.com>.



1998] END-OF-LIFE DECISION MAKING 21

matter."*^ Consistent with the jurisprudence of the Rehnquist Court, the Supreme
Court simply endorsed the state's authority either to maintain or to change the

law largely without constraint from the federal courts."*^ The Rehnquist opinion
relied on historical state law prohibitions against aiding and abetting suicide'*^

and accorded great deference to the state's assessment of potential harms under

the rational basis test.^° The Ninth Circuit, without the deference exercised by
the Supreme Court, more substantively addressed arguments for and against
legalization.^^ Therefore, the structure of its responses is most likely to reappear

on the state level in the ensuing debate.

In addition to differing on the level of deference required, the Supreme
Court's and the Ninth Circuit's majority opinions differed on the sources they

relied upon as persuasive. To the extent that the Rehnquist opinion did engage
the arguments for and against legalization substantively, it relied largely on the

report of the New York State Task Force on Life and the Law." In this, the
Rehnquist opinion resembled state court decisions of the last two decades which

relied heavily on the recommendations of the President's Commission to
recognize a right to refuse or forego life-sustaining treatment.^^

In contrast, in major parts of its majority opinion, the Ninth Circuit addressed

the question of the legal status of assisted suicide with more of an emphasis on
empirical evidence and its relevance for resolving public policy and legal issues.

In other parts of its decision, the Court of Appeals made assumptions about
professional behavior that make the question of legalization appear to rest on
empirically verifiable facts. The Ninth Circuit's opinion is quite revealing in its

treatment of arguments that rely on apparently empirically-based or empirically

verifiable assertions, a point that is especially pertinent to the question presented

by this special issue of the Indiana Law Review.
This critical analysis ofthe impact of "empiricalism" on the Ninth Circuit's

decision is not intended to illustrate deficiencies unique to that opinion. An
examination of the earlier Ninth Circuit opinion written for a panel of the Ninth

Circuit by Judge John Noonan^"* and vacated by the Court of Appeals en banc in

Compassion in Dying, for example, would likely yield a similar critique even

though Judge Noonan's opinion reached a result opposite to that of the later

majority. No matter the side, arguments for and against legalization of assisted
suicide, to the extent they engage the substantive issues, present similar

questions regarding the availability, reliability and relevance of empirical data

on public and professional opinion, current practices, and anticipated risks and

47. Washington v. Glucksberg, 1 17 S. Ct. 2258 (1997).

48. See id. at 2275.

49. Id. at 2262-65.

50. Mat 2271-75.
5 1

.

Compassion in Dying v. Washington, 79 F.3d 790, 816-39 (9th Cir. 1 996).

52. Glucksberg, 1 1 7 S. Ct. at llll-lS.

53. See supra notes 27-32 and accompanying text.

54. Compassion in Dying v. Washington, 49 F.3d 586 (9th Cir. 1995), vacated en banc, 79

F.3d 790 (9th Cir. 1996), rev 'd sub nom. Washington v. Glucksberg, 1 17 S. Ct. 2258 (1997).



22 INDIANA LAW REVIEW [Vol. 31:13

benefits of the rejection or maintenance of the current legal status of assisted

suicide.

The debate over physician-assisted suicide embraces many arguments. Three
of several identifiable arguments in this debate are:

(1) The majority of Americans appear to indicate some support for
physician-assisted death,^^ and therefore: law must change to reflect public

opinion and actual practice;^^ or current law^ actually does accurately reflect

society's "real" opinion concerning legalization,^^ if majority views are even

relevant to the question.
^^

(2) The participation of doctors in the process of assistance in suicide, as
compared to the participation of family members for example, alters the nature
of the act: participation of doctors contributes to the safety of the process;^^ or

the participation of doctors merely "medicalizes" or "white coats" the act so that

it appears to be safer, more normal, and more acceptable.^°

(3) Particular "vulnerable" populations may be affected differently by the
legalization of assisted suicide: legalization will not harm these populations and

may remedy the special suffering faced by vulnerable populations;^* or
legalization presents a special danger within a system that already victimizes

certain identifiable populations.^^

55. See James A. Tulsky et al., A Middle Ground on Physician-Assisted Suicide, 5

Cambridge Q. Healthcare Ethics 33, 34 (1995) (focusing also on the consistent refusal to indict,

prosecute or convict physicians who aid in death); see also David Orentlicher, Physician

Participation in Assisted Suicide, 262 JAMA 1 844 ( 1 989).
56. See, e.g., Catherine L. Bjorcic, Physician-Assisted Suicide: Whose Life Is It Anyway, 47

SMUL. Rev. 371, 372 (1994) (arguing that opinion surveys, suicide statistics and public concern
demand legalization of physician-assisted suicide).

57. Brief of the United States Catholic Conference et al. as Amici Curiae in Support of

Petitioners at 8 n.l, Washington v. Glucksberg, 1 17 S. Ct. 2258 (1997) (No. 96-1 10); see also

Stephanie Graboyes-Russo, Too Costly to Live: The Moral Hazards ofa Decision in Washington

V. Glucksberg and Wacco v. Quill, 51 U. MIAMI L. REV. 907, 915-16 (1997) (arguing that polls only

reflect attitudes toward dying with dignity not toward physician-assisted suicide).

58. Edmund D. Pellegrino, The Limitation ofEmpirical Research in Ethics, 6 J. CLINICAL

Ethics 161, 162 (1995); see also Ruth Shalit, When We Were Philosopher-Kings: The Rise ofthe

Medical Ethicist, NEW REPUBLIC, April 28, 1997, at 24 (criticizing bioethicists and the "place of
ethical expertise in a liberal democracy").

59. See, e.g., Lori D. Pritchard Clark, Rx: Dosage ofLegislative Reform to Accommodate

Legalized Physician-Assisted Suicide, 23 CAP. U. L. REV. 689, 705 (1994).

60. See John M. Dolan, Is Physician-Assisted Suicide Possible?, 35 DUQ. L. REV. 355, 392-

93 (1996); see also George J. Annas, Physician-Assisted Suicide—Michigan 's Temporary Solution,
20 OhioN.U. L. Rev. 561, 569 (1994) (arguing that physicians should not sacrifice their ethics to

further the goals of the state).

61

.

Brief of Council for Secular Humanism and International Academy of Humanism as

Amici Curiae in Support of Respondent at 9-10, Washington v. Glucksberg, 1 17 S. Ct. 2258 (1997)

(No. 96-110).

62. See, e.g., Alexander M. Capron, Legalizing Physician-Aided Death, 5 CAMBRIDGE Q.



1998] END-OF-LIFE DECISION MAKING 23

Each of these three mirror-image points either relies on empirical data or
suggests that empirical data will be relevant to the resolution of the dispute over

the legalization of physician-assisted suicide. For example, the Ninth Circuit

majority opinion in Compassion in Dying used public opinion polls as relevant

evidence supporting judicial recognition of a constitutional right to physician-

assisted suicide.^^ The court relied on the protective role of physicians as an
essential characteristic of the legal right it recognized.^ Further, the court

rejected most claims that a legal right to assistance in suicide will operate in

unacceptable population-specific patterns and argued that the availability of

legalized physician assistance in suicide may address the particular needs of
some persons within these groups.^^

A. Opinion Polls

The Ninth Circuit described broad public support for the legalization of
physician-assisted suicide. The court began its discussion of public opinion by
observing that "[pjolls have repeatedly shown that a large majority of
Americans—sometimes nearing 90%—fully endorse recent legal changes
granting terminally ill patients, and sometimes their families, the prerogative to

accelerate their death by refusing or terminating treatment.'*^ The court further
reported that "[o]ther polls indicate that a majority of Americans favor doctor-

assisted suicide for the terminally ill,"^^ and described an April 1990 Roper poll

that found "64% of Americans believed that the terminally ill should have the
right to request and receive physician aid-in-dying.'** The court also described
"another national poll" that showed "nearly two out of three Americans favor
doctor-assisted suicide and euthanasia for terminally ill patients who request it"^^
and stated that a 1994 Harris poll found "73% of Americans favor legalizing
physician-assisted suicide."^*'

Healthcare Ethics 10, 19-20 (1996) (discussing the particular vulnerability women); Marshall

B. Knapp, Old Folks on the Slippery Slope: Elderly Patients and Physician-Assisted Suicidey 35

DUQ. L. Rev. 443, 444 (1996) (arguing that the elderly are a particulary vulnerable group

susceptible to influence on physician-assisted suicide by family members and physicians because

they may believe themselves a burden to their family or society).

63. See discussion infra Part LA.

64. See discussion infra Part LB.

65. See discussion infra Part I.C.

66. Compassion in Dying v. Washington, 79 F.3d 790, 810 (9th Cir. 1995) (citing Sanford

H. Kadish, Letting Patients Die: Legal and Moral Reflections, 80 Cal. L. Rev. 857, 860 & n.l6
(1992)).

67. Id

68. Id. (citing Robert L. Risely, Voluntary Active Euthanasia: The Next Frontier, Impact

on the Indigent, 8 ISSUES IN L. & Med. 361, 365 (1992)).
69. Id. (quoting Kadish, supra note 66, at 861 n.22).

70. Id. (citing David Cannella, Physician-Aided Suicide Fight Rages in Several States:

Issue Expected to Go to Supreme Court, ARIZ. REPUBLIC, May 13, 1995, at A22).



24 INDIANA LAW REVIEW [Vol. 31:13

In the same paragraph with its recitation of the results of public opinion
polls, the court also described the voting margins on the physician-assisted

suicide referenda and state initiatives in Oregon, California, and Washington.^^

In the defendant state, Washington, the proposed legislation produced only 46%
affirmative votes;^^ however, in Oregon, proposed legislation to legalize

physician-assisted suicide had passed with 51% of voters approving the
referendum 7^

The court concluded its description of public opinion, including the results
ofthe California, Washington and Oregon initiatives on legalization of assisted
suicide, by concluding that "there is unquestionably growing popular support for

permitting doctors to provide assistance to terminally ill patients who wish to
hasten their deaths."^"* The opinion implied that the degree of public support for
legalization is relevant to its resolution of the constitutional issueJ^ Growing
approval could be used to indicate that its holding that there is a constitutional

right to assistance in suicide reflects mainstream thought.

The Ninth Circuit's use ofpublic opinion polls concerning physician-assisted
suicide raises at least two issues that are specific to the suicide debate. First,

although the Ninth Circuit considered public opinion to be relevant to the

recognition of a liberty interest in physician assistance in death, it did not

confront the gap between public opinion polls and the results of public votes on

the issue to that point. Second, the court did not recognize that studies of both

public and professional attitudes and practices relating to aid in dying are

confounded by the transformation occurring in the language of end-of-life

decisions.

The court described the results of the three public initiatives/referenda in the
same paragraph as it discussed public opinion polls and concluded that the
electoral results provide evidence of growing public support for legalization of

assisted suicide. The court's reliance on the failed referendum in Washington
state, the defendant in Compassion in Dying, as support for its holding that there

is a constitutional right to assistance in suicide is curious. The court interpreted
the clearly negative vote in Washington as positive support of its own position.^^

The Ninth Circuit decision used evidence of public opinion polls to support
its rejection of the results of a direct vote on the statute at issue in the case.^^

Public referenda may be irrelevant to any particular individual civil liberty,^^ of

71. Id.

72. Id.

73. Id

74. Id. In November 1997, Oregon voters voted not to repeal their earlier affirmative vote

on assisted suicide.

75. /^. at 811.

76. Mat 810.
77. Id

78. For a discussion of the role of popular votes on bioethical issues, see Judith F. Daar,

Direct Democracy and Bioethical Choices: Voting Life and Death at the Ballot Box, 28 U. MiCH.

J.L. Reform 799 (1995).



1 998] END-OF-LIFE DECISION MAKING 25

course, and the Ninth Circuit certainly did view access to physician-assisted

death as a liberty interest that reached the most intimate decisions7^ The Ninth
Circuit, however, argued that public opinion polls were relevant to the

recognition of this right. The court reviewed the jurisprudence on "fundamental

rights" and noted the consistent effort to assure that the recognition of

fundamental rights did not emerge only from the "imposition ofthe Justices' own
choice of values."^^ The use of public opinion polls in constitutional

adjudication is controversial. Generally, public opinion polls have not been

persuasive when courts were dealing with issues of fundamental constitutional
rights. Opinion polls have been labeled as being too uncertain because they are

based on temporary opinions which are continually swayed by political

activists.^' The United States Supreme Court, considering a death penalty case,
believed that opinion polls were inconclusive because the people had not voted

on the issue through referendum or through their representatives in the

legislature, which the court believed would be done if a position was truly
popular.^^ Recently, a current Supreme Court Justice has spoken on the issue and

submitted that the Federal Judiciary was given Article III protections (life tenure

and salary protection) under the United States Constitution so that it could decide

fundamental constitutional issues independent of such things as opinion polls.^^

Despite its own discussion of public opinion, the Ninth Circuit finally
specifically rejected both majority and minority control of this issue: "[N]either

the state nor the majority of the people in a state can impose its will upon the

individual" in this matter.^"* The court rejected arguments that the issue should

be left to state elections, citing legal conflicts that may arise if some states
legalized assisted death and others did not.^^

Commentators have offered other reasons for rejecting the Washington and

California votes as invalid indicators of public sentiment. For example, after the

failed initiatives in California and Washington, commentators charged that the

vote did not reflect the actual will of the majority because certain groups

overwhelmed the proponents of legalization in terms of money spent on the
campaign.^^ Others implied that certain "outside" groups heavily influenced the

79. Compassion in Dying, 79 F. 3d at ^]2-\3.

80. Id at 803.

81. Swann v. Charlotte-Mecklenburg Bd. of Educ, 318 F. Supp. 786, 793 (W.D.N.C. 1970).

82. Stanfordv. Kentucky, 492 U.S. 361,377(1989).

83

.

Hon. Justice Clarence Thomas, Judging, 45 U. Kan. L. Rev. 1,4(1 996) (Justice Thomas

argued that ifjudges allow their "decisions to be guided by popular sentiment and group rights and

demands, then the Constitution will be nothing but a malleable, transparent barrier to majoritarian

desires.").

84. Compassion in Dying, 79 F.3d at 839.

85. /^. at 833 n. 124.

86. Diane M. Gianelli, Euthanasia Measures Fail, But Backers Vow to Renewed Push, AM.

Med. News, Nov. 23, 1992, at 42; Susan Gilmore, Will Foes' Efforts Doom California's
Euthanasia Bill?, SEATTLE TIMES, Oct. 24, 1992, at A14.



26 INDIANA LAW REVIEW [Vol. 31:13

state-specific public debate.^^ Some advocates charged that opponents
sensationally misrepresented the potential impact of the proposed legislation.^^

Others conceded that the legislation could have been rejected on the basis of

specifics—^the weakness of safeguards, for example—even though there may
have been adequate support for more narrow legalization.^^ Still, some
commentators have observed that pre-election polls may indicate that public
support is strong enough to win as they did in Washington, but cannot indicate,

until the actual vote, whether support is deep enough; that is, whether it will

persist once arguments against the initially preferred position are offered or once

an actual vote or other decision is required.^^

The second issue in the court's use of survey data concerning acceptance of
physician-assisted suicide relates to the emerging problem of language in life-

sustaining treatment issues. The suicide debate, like some other politicized
bioethics issues,^' is plagued by jockeying over the names that will identify the
advocate's position with virtue and caregiving and the opponent's position with

evil and danger. In part, this is an unavoidable consequence of the politicization

ofthe question: the battle takes place in snippets and slogans and has to appeal

to emotion and intuition through association with familiar terms.

The language strategy may pay off politically. In 1993, in preparation for the
referendum in Oregon, a survey was taken to test the impact of different terms.^^

In that survey, 44% of respondents indicated that they would vote for a law
allowing "physician-aided suicide"; 51% for "physician's aid in dying"; 55% for

87. Jim Simon, National Groups, Huge Budgets Invade Former Turf of the Little Guy,

Seattle Times, Oct. 4, 1 99 1 , at A 1

.

88. Ellis E. Conklin, Support for Initiative 119 Slipped Away in Final Hours, SEATTLE

POST-INTELLIGENCER, Nov. 7, 1991, at A9; Tom Paulson, Death With Dignity Backers Accuse Foes

ofDistortion, SEATTLE POST-lNTELLIGENCER, Oct. 24, 1991, at Bl.

89. Howard Breuer, Act ofMercy Brings Legal Problems Santa Paula Man Probed in Ailing

Wife 's Suicide, L.A. DAILY NEWS, Oct. 17, 1993, at TOl; B.D. Colen, Campaign '92 Initiatives

California is Voting on the Right to Die Act Would Legalize Some Forms of Mercy Killing,

Newsday, Oct. 29, 1992, at 23.

90. See Ezekiel J. Emanuel, Empirical Studies on Euthanasia and Assisted Suicide, 6 J.

Clinical Ethics 158 (1995); Gianelli, supra note 86, at 42; Conklin, supra note 88, at A9. The

Ninth Circuit commented that the Washington and California proposals "contained far fewer

practical safeguards" than did the "carefully-crafted" referendum in Oregon. Compassion in Dying,

79 F.3d at 810. Later in its opinion, however, the court rejected a central restriction in the Oregon

referendum. The court implied in dictum that restricting assisted suicide to prescription of

medication which the patient would administer to himself or herself may itself be unconstitutional.

/^. at 831-32.

91

.

Another example is the debate over "partial-birth abortions." See Kathleen A. Cassidy

Goodman, The Mutation ofChoice, 28 St. Mary's L.J. 635 (1997); Kim Painter, Handling the

Aftermath ofthe Decision to Abort, USA TODAY, Aug. 18, 1977, at 6D; Richard Saltus, Late-term
Abortion at Issue Called Rare in Mass., BOSTON GLOBE, May 17, 1997, at Al.

92. Mark O'Keefe, Doctor-Assisted Suicide Law Boils Down to a War of Words, THE

Oregonian, Dec. 1, 1994, at Al.



1 998] END-OF-LIFE DECISION MAKING 27

"euthanasia"; and nearly 66% for a terminally ill patient's choice to "die with
dignity."^^ The survey revealed an increase of 50% in affirmative responses as
between "physician-aided suicide" and "death with dignity" just by use of the

different term.^"^ Equally revealing is the difference between "physician-aided

suicide" and "physician's aid in dying":^^ the difference between 44% and 51%
is the difference between winning and losing an election. The Ninth Circuit's
use of polls asking about "aid in dying" as evidence of support for physician-

assisted suicide demonstrates a lack of recognition of these language issues.^^

Although the term "physician-assisted suicide" connotes professional

assistance to an individual engaging in an act, committing suicide, that is not

itself illegal, proponents of legalization no longer rely on the assisted "suicide"

concept or language. The campaign director for the 1992 California initiative

reacted to the campaign's victory in eliminating the word suicide from the ballot

by stating that "the ballot language is worth a million bucks to us.'*^

The Oregon initiative itself provides that "Actions taken in accordance with

this Act shall not . . . constitute suicide [or] assisted suicide."^^ The Ninth Circuit

explicitly challenged the use of the "s" word to describe the intervention ^^ even

though the opinion laboriously examined the historical treatment of suicide to

prove that suicide was not always viewed as immoral or illegal.
'°°

The political power of language predictably has led advocates to search for

a term for "assisted suicide" or "aid in dying" or "lethal intervention" that

communicates a certain quality about the interaction of physician and patient.

The language strategy will have an impact on the political process if politics
holds true to form. But strategic language has produced a more substantial
difficulty in this case. At a minimum, it has had an impact on the evaluation and
meaning of empirical data describing current opinions and practices relating to

end-of-life care. The poll discussed earlier,'^^ which tested only differences in
the terms, illustrates how powerfully the terminology can affect the results of
surveys probing this issue.

The political use of language is not the only difficulty here. Underneath the
political currency lies a fundamental deconstruction of the ethical and legal

framework established for decisions regarding care at end of life. An essential
point in the debate over assisted suicide is whether there is any difference at all

between actions that are currently illegal and those that are legal and widely

accepted. Advocates for legalization of assisted suicide argue that withholding

93. Id.

94. Id.

95. Id

96. Compassion in Dying, 79 F.3d at 810.

97. Paul Jacobs, Prop. 161—A Matter ofLife or Death at the Polling Place, L.A. TIMES,
Oct. 10, 1992, at 20.

98. Or. Rev. Stat. § 127.880 (1996) (Oregon Death with Dignity Act § 3.14).

99. Compassion in Dying, 79 F.3d at 824.

100. /£/. at 806-10.

101. 5ee O'Keefe, 5wpra note 92.



28 INDIANA LAW REVIEW [Vol. 31:13

or foregoing life sustaining treatment (particularly nutrition and hydration, but

also ventilator support and other interventions) is no different than providing

lethal medication. ^^^ Proponents of legalization also argue that administering

medication to relieve pain, where the pain medication may foreseeably hasten the
patient's death, is no different than providing a non-therapeutic lethal drug.'^^

Withdrawal of medical treatment is legal within broadly circumscribed

circumstances while providing or administering a lethal drug is not. Prescribing

or administering pain medication therapeutically, even assuming that the

medication may hasten death, which itself is a highly controversial point,^^ is not
illegal. '^^ Opponents of legalization argue that there is an ethical, historical,

medical, and legal difference between aiding suicide and withdrawing treatment

or treating pain.^^^ The Ninth Circuit itself recognized that both withdrawal of
treatment and provision of adequate pain relief are legal. ^°^ The Ninth Circuit
adopted the argument that there is no legally significant difference between those

actions and providing lethal medication, while the Supreme Court, in contrast,
detailed the states' consistent adherence to legal distinctions between

withdrawing treatment and medicating for pain on the one hand and providing

lethal medication on the other. ^^^ The Rehnquist opinion and the concurring
opinions further emphasized the legality of providing pain medication even

where that medication may present the risk of an earlier death.^°^
Establishing a distinction in legal and ethical character between withdrawal

oftreatment and euthanasia or assisting suicide was a critical point in the courts'
initial legal recognition of a patient's right to refuse medical treatment.

*^°

102. See, e.g.. Brief of Julian W. Whitaker, M.D., as Amicus Curiae in Support of

Respondents at 15, Washington v. Glucksberg, 117 S. Ct. 2258 (1997) (Nos. 96-110, 95-1858)

[hereinafter Whitaker]; Brief ofGay Men's Health Crisis and Lambda Legal Defense and Education

Fund as Amicus Curiae in Support of Respondents at 24, Washington v. Glucksberg, 1 17 S. Ct.

2258 (1997) (Nos. 96-1 10, 95-1858).

103. See Whitaker, supra note 102.

104. See, e.g., John Colin Partridge & Stephen N. Wall, Analgesiafor Dying Infants Whose
Life Support is Withdrawn or Withheld, PEDIATRICS, Jan. 1, 1997, at 76 (reporting a study

indicating that average length of time between withdrawal of ventilator support and death was

identical for infants receiving no pain medication and those receiving morphine and longer for those

infants receiving a greater amount of morphine).

1 05. Yale Kamisar, The "Right to Die ": On Drawing (and Erasing) Lines, 35 DUQ. L. Rev.

481,500(1996).

106. Brief Amicus Curiae in Support of Petitioners at 3, Vacco v. Quill, 1 17 S. Ct. 2293

(1997) (No. 95-1858); Brief of the United States Catholic Conference et al. as Amici Curiae in

Support of Petitioners at 4-5, Vacco v. Quill, 1 17 S. Ct. 2293 (1997) (No. 95-1858).

107. Compassion in Dying v. Washington, 79 F.3d 790, 827 (9th Cir. 1996).

108. Washington v. Glucksberg, 1 17 S. Ct. 2258, 2270 (1997).

109. Id. at 2270; id. at 2276 (Souter, J., concurring); Washington v. Glucksberg, 1 17 S. Ct.

2302, 2307 (O'Connor, J., concurring).

1 1 0. See, e.g., Alexander M. Capron, Borrowed Lessons: The Role ofEthical Distinctions

in Framing Law on Life-Sustaining Treatment, 1984 ARIZ. ST. L.J. 647.



1998] END-OF-LIFE DECISION MAKING 29

Persuading health care professionals that there was a qualitative ethical
difference between allowing a patient to die and killing a patient was also
important in the effort to assure that withdrawal of life-sustaining treatment and

effective pain management would be carried out.''^
Whether or not there is a "real" difference between withdrawing treatment

or medicating for pain on the one hand and assisted suicide or euthanasia on the

other, the suicide debate's deconstruction of the legal and ethical underpinnings

ofthe last two decades (and along with that foundation, the language) has made
it difficult to communicate. A prominent example of this difficulty emerged
from a recent survey of critical care nurses.^ *^ This survey was described as
revealing that nineteen percent of the nurses had reported performing or

participating in "euthanasia," with the implication that the nurses had engaged

in illegal or unethical behavior, that they had killed their patient."^ This was not
just another incident where the media could be criticized for misunderstanding

or hyperbole. The study's principal investigator, Dr. David Asch, described his
survey as supporting that finding.""*

The publication of the survey elicited a strong response. Much of the
negative reaction was heavily critical ofthe survey questions. Two questions in
particular were controversial.

The survey asked whether the nurse respondent had ever performed actions
"with the intent of causing or hastening that patient's death—other than the
withdrawal of life-sustaining treatment" and whether the nurse had ever given an

"overdose of opiates.""^ Both of these questions alluded to the administration

of medication to a person whose death is imminent or who is in the final stages
of terminal illness. In response to the publication of the Asch study, nurse
specialists, professional nursing organizations, and nurse researchers generally

claimed that the questions asked on the survey were too ambiguous to generate

a meaningful response."^

The question referring to an "overdose of opiates" is particularly
problematical. "Overdose" implies that the pain medication was administered in
an amount that was not required for the treatment of pain but that was instead

HI. Graboyes-Russo, supra note 57, at 9 1 6- 1 7.

112. David A. Asch, The Role ofCritical Care Nurses in Euthanasia and Assisted Suicide,

334 New Eng. J. Med. 1374 (1996).
113. Barbara Reynolds, Nurses, Physicians ShouldNot Play God, USATODAY, June 7, 1 996,

at 15A; Are Nurses Angels ofDeath?, TIMES UNION, June 1, 1996, at A6.

114. David A. Asch, Promises and Pitfalls Along the Road to Empirical Scholarship in

Bioethics, 2 CENTER FOR BlOETHICS NEWSLETTER 1 ( 1 996).

115. David A. Asch, The Role ofCritical Care Nurses in Euthanasia and Assisted Suicide,

335 New Eng. J. Med. 971, 973-74 (1996) (Letter to the Editor).
1 1 6. Colleen Scanlon, Euthanasia and Nursing Practice—Right Question, Wrong Answer,

334 New Eng. J. Med. 1401 (1996); see also Patricia A. Dunn, The Role ofCritical Care Nurses
in Euthanasia and Assisted Suicide, 335 NEW ENG. J. MED. 971 (1996) (Letter to the Editor);
Nancy L. Szaflarski & John M. Clochesy, The Role ofCritical Care Nurses in Euthanasia and
Assisted Suicide, 335 NEW Eng. J. MED. 971 (1996) (Letter to the Editor).



30 INDIANA LAW REVIEW [Vol. 31:13

non-therapeutic. Current research on effective pain relief has clearly established

that "standard" doses of opiates are meaningless for patients who have received
these pain medications over some amount of time.' '^ Much larger doses of the
drugs are both safe for such patients and are required for relief of pain.* '^ Such

"overdoses" can be well tolerated by those patients often without the sedation

side effect or other side effects. For minimally adequate treatment of pain in

patients who have used prescribed pain medication previously, doctors should be
prescribing and nurses administering "overdoses" of opiates and can prescribe

an "overdose" without shortening the patient's life."^ If Asch intended to
exclude these totally therapeutic, non-life-threatening "overdoses" from the reach

of his question, he failed to do so. If he intended to include any case in which a

"large" amount of opiates is provided, he has asked a question using a term that

implies substandard care to describe acceptable medical or nursing care.

Asch responded to the criticism of his study:

[T]he term "euthanasia" is loaded. Those who believe that the term can
refer properly only to activities that are immoral may also feel that it
cannot apply to all the activities reported by the nurses. If there is a

continuum of moral appropriateness represented here, it is not clear

where the moral divide lies, whether there is a single divide, or whether

that divide is shifting over tim^. The range of activities described by the

nurses who participated in this study may reveal the inadequacy of the
term "euthanasia" and the many professional and legal policies built on
it.'2^

Asch's response highlights the ambiguity of the term euthanasia but it also

illustrates the difficulty of the context in which the debate over and studies of

physician-assisted suicide are occurring. Whether or not the argument that there

is no difference, other than the current legal distinction, between withdrawing

treatment and medicating for pain on one hand and assisting in suicide on the

other ultimately prevails, the "no distinction argument" has a direct impact on the

meaningfulness of empirical studies on professional attitudes and practices. To
the extent that such research may be used in judicial or legislative decision
making, this problem has to be recognized specifically and with some
sophistication.'^*

The "no distinction argumenf could be having other broad and disturbing
consequences. The desired outcome of the deconstruction of the current ethical
and legal framework may be the expansion of medical options and the expansion
of individual control over medical decisions, as well as the relief of suffering.

117. See, e.g., Russell K. Portenoy, Opioid Therapy for Chronic Nonmalignant Pain:

Clinicians ' Perspective, 24 J.L. MED. & ETHICS 296 (1996).
118. Id.

1 19. See, e.g., Szaflarski & Clochesy, supra note 1 16.
120. Asch, supra note 1 1 5, at 973.

121. For a critical analysis of survey data concerning physician attitudes toward euthanasia

and assisted suicide, see Emanuel, supra note 90, at 158.



1 998] END-OF-LIFE DECISION MAKING 3

1

It is possible, however, that dismantling the legal and ethical framework for the
treatment of pain could have an opposite effect. If medicating for pain relief

becomes associated with physician-assisted suicide or euthanasia, adequate
treatment for pain could take on all of the moral connotations of killing.

Individual doctors and nurses who believe that it is immoral to "kill" a patient
and who may already be anxious about providing adequate pain medication for
their patients for fear of causing death, may become even more hesitant to
provide medically required pain medication.^^^ This reaction would exacerbate
the current shameful neglect of pain and cause increased and avoidable human
suffering. Furthermore, legislators who want either to make assisted suicide and
euthanasia illegal, or to legalize assisted suicide and euthanasia but only under

tightly controlled circumstances, might establish standards that greatly restrict

access to effective pain relief For example, in the 1997 legislative session, a bill

was filed in the Michigan legislature that would legalize physician-assisted
suicide.

^^^ A section of that bill, however, provided: "A nonphysician . . . who
administers [or] causes to be administered . . . medications ... to an individual

for use in a manner substantially likely to cause or hasten that individual's death

is guilty ofa felony."'^"* This provision would prohibit nurses from administering

pain medication to patients who are suffering severe pain if the medication
carries the risk of hastening death. Doctors prescribe medication, but nurses

ordinarily administer medications in a hospital setting. Nurses ordinarily attend

patients at the end of life. Physicians are often not readily available to administer

pain medication as needed. This proposed legislation, then, legalizes the use of

medication to assist a death when certain restrictive circumstances are met but
could make the use of pain medication illegal where the narrow circumstances
of the statute are not met.

B. How Physicians Do and Will Behave

Both Vacco v. Quill and Compassion in Dying limited their holdings to

assistance provided by physicians. The three state initiatives placed before
voters in California, Washington and Oregon also legalized assistance in suicide

only where it was to be provided by physicians. '^^

Opponents, as well as some proponents, of legalization of assisted suicide
have taken issue with the use of physicians to aid in death. Some opponents of
legalization of assisted suicide argue that the involvement of physicians perverts

the role of physician as healer and will irreparably damage the trust in the
physician-patient relationship. ^^^ Other opponents of legalization argue that

122. Melissa L. Buchan & Susan W. Tolle, Pain Relieffor Dying Persons: Dealing with
Physicians ' Fears and Concerns, 6 J. CLINICAL ETHICS 53, 57 (1995).

123. S. 653, 89th Leg., Reg. Sess. (Mich. 1997).

124. /c/. §5689.

125. Washington v. Glucksberg, 1 17 S. Ct. 2258, 2266 (1997); see also Compassion in Dying

V. Washington, 79 F.3cl 790, 810 (9th Cir. 1996).

126. Brief of the American Medical Association et ai. as Amici Curiae in Support of



32 INDIANA LAW REVIEW [Vol. 31:13

reliance on physicians as the centerpiece of the movement to legalize assistance
in death is strategic to make the process appear to be an ordinary form of medical
treatment. '^^ Some proponents of legalization argue over whether doctors should
have a monopoly over this work and whether other professionals, including
nurses with prescribing authority or newly created categories of health care
providers, would be better suited to the task.'^^ Other proponents argue that the
patient should be able to choose to have the assistance of family or friends

instead of that of a professional stranger, and that legally requiring any

professional intervention is an invasion ofthe privacy ofthe patient and gives too

much control to physicians. '^^

The Ninth Circuit emphasized the protective and tempering role of
physicians in the provision of assistance in suicide. It viewed the participation

of doctors as a safeguard against abuse:

We believe that most, if not all, doctors would not assist a terminally ill
patient to hasten his death as long as there were any reasonable chance

of alleviating the patient's suffering or enabling him to live under
tolerable conditions. We also believe that physicians would not assist
a patient to end his life if there were any significant doubt about the

patient's true wishes. To do so would be contrary to the physicians'
fundamental training, their conservative nature, and the ethics of their

profession.
'^°

In this statement, the court attributes two priorities to physicians in the care

oftheir patients at the end of life. First, the court alludes to doctors' commitment

to relieving suffering and pain. Second, the court emphasizes physicians'

adherence to the wishes of their patients. The court nests these priorities in
medical education and medical ethics.

No one can know how physicians will behave if assistance through lethal
intervention were legalized. It would seem likely that neither Jack Kervorkian

nor Timothy Quill—^the two most prominent physician-advocates of physician-
assisted suicide—^would represent the majority of physicians providing such
assistance. There is empirical evidence, however, that examines physicians'

Petitioners, Washington v. Glucksberg, 117 S. Ct. 2258 (1997) (No. 96-1 10); Brief of the Legal

Center for Defense of Life, Inc. and the Pro-life Legal Defense Fund as Amici Curiae in Support

of Petitioners, Washington v. Glucksberg, 1 17 S. Ct. 2258 (1997) (No. 96-1 10).

127. Brief of the American Medical Association et al. as Amici Curiae in Support of

Petitioners, Vacco v. Quill, 1 17 S. Ct. 2293 (1997) (No. 95-1858); Brief of the American Medical

Association et al. as Amici Curiae in Support of Petitioners, Washington v. Glucksberg, 1 17 S. Ct.

2258 (1997) (No. 96-110).

128. See, e.g., Annette E. Clark, Autonomy and Death, 71 TULANE L. REV. 45 (1996).

129. David C. Thomasma, An Analysis of Arguments For and Against Euthanasia and

Assisted Suicide: Part One, 5 CAMBRIDGE Q. HEALTHCARE ETHICS 62, 66-67 (1996) (describing

claims that there should be no "middle person" involved and evidence of physician paternalism in

refusals to comply with euthanasia requests).

130. Compassion in Dying, 79 F.3d at 827.



1 998] END-OF-LIFE DECISION MAKING 33

behavior in their care of patients generally, particularly in their treatment of

patients in pain, and in their compliance with patients' choices concerning

medical treatment.

The Ninth Circuit's confidence in the medical commitment to providing
effective pain relief through available means is not supported in currently
available data. The medical capacity to relieve pain is greater than it has ever
been. Effective pain management is medically available for cancer pain, for
chronic nonmalignant pain, and for pain related to diseases at the end of life.

Effective pain management is available in many forms, including controlled
substances such as opioids, that do not cause addiction or serious mental

impairment.
^^^

Yet, pain continues to be seriously neglected and undertreated. Treatable but

untreated pain is a widespread problem that cuts across many patient populations.
Studies have repeatedly documented undertreatment of pain in U.S. health care.

For example, the SUPPORT study of 9105 patients dying in five teaching
hospitals found pain management lacking. SUPPORTroportQd that the surviving
family members of fifty percent of the dying patients reported that the patients
suffered moderate to severe pain half of the time.'^^ The very recent Institute of
Medicine report on end-of-life care also identified undertreatment of pain as a

major concern. '" Seventy-five percent of cancer patients in one study reported

suffering pain, with forty to fifty percent reporting moderate to severe pain and

twenty-five to thirty percent reporting severe pain. This occurs even though

ninety percent of cancer pain can be relieved through "relatively simple

means."'^'^ Chronic nonmalignant pain has been described as an extremely

prevalent problem. '^^ Over two-thirds of nursing home residents experience
serious chronic pain.'^^ Moreover, the elderly, minorities, women, children, and
those unable to speak for themselves due to disability bear the brunt of

ineffective care and are undertreated at even higher rates than others. ^^^ Despite

131. See Agency for Health Care Policy & Research, U.S. Dept. of Health & Human
Services, Management of Cancer Pain: Management of Cancer Pain Guideline Panel

(1994) (Pub. No. 94-0592); AGENCY FOR HEALTH CARE Policy & Research, U.S. Dept. of
Health & Human Services, Acute Pain Management: Operative or Medical Procedures
AND Trauma ( 1 992) (Pub. No. 92-0032).

132. SUPPORT, supra note 25, at 1591.

133. Institute of Medicine, Approaching Death: Improving Care at the End of Life

(Marilyn J. Field & Christine K. Cassel eds., 1997).
1 34. Agency for Health Care Policy & Research, supra note 1 3 1 (referring generally

to both articles).

135. Russell K. Portenoy, Opioid Therapyfor Chronic Nonmalignant Pain: A Review ofthe

Critical Issues, 1 1 J. PAIN «fe SYMPTOM MGMT. 203 (1996).

136. Bruce A. Ferrell et al., Pain in the Nursing Home, 38 J. AM. GERIATRICS SOC. 409

(1990).

1 37. B. S. Shapiro & B. R. Ferrell, Pain in Children and the Frail Elderly, in APS BULLETIN
(Oct./Nov. 1992); Knox H. Todd et al.. Ethnicity as a Risk Factor for Inadequate Emergency

Department Analgesia, 269 JAMA 1 537 ( 1 993).



34 INDIANA LAW REVIEW [Vol. 31:13

the development of effective pain management interventions and the overall
human and financial cost of pain, pain is neglected and undertreated.

The Ninth Circuit also expresses remarkable confidence in doctors' interest
in and commitment to following the patient's wishes concerning medical
treatment at the end of life. The Ninth Circuit's confidence greatly exceeds what
the empirical data would support.

*^^

In an important article published in the Journal of the American Medical
Association, David Orentlicher, then writing for the Office of General Counsel

of the AMA, and now co-director of Indiana University's Center for Law and
Health, observes:

Over the past two decades a societal consensus has developed around the
principle that decisions about life-sustaining treatment should be guided

by patient self-determination. According to the President's Commission,

the Hastings Center, the American Medical Association, and the U.S.

Supreme Court, treatment decisions should be based on the values, goals

and preferences of the patient.

While theory may emphasize the patient's values, empirical data suggest
that other considerations may have a greater impact on decisions about
life-sustaining treatment. In particular, there is increasing evidence that

physician values may be a more decisive factor than patient values in
these decisions.

'^^

Orentlicher supports his statement with a review of the empirical literature.

For example, the authors of a study of compliance with advance directives

concluded that the physicians in the cases studied actually provided undesired

treatment and withheld desired treatment based on the physician '5 judgment of

what was appropriate or beneficial to the patient. Orentlicher concludes that
"patient's preferences were respected as long as the physicians thought that the

patients' choices resulted in the best decisions."^'*^

Studies also consistently indicate that patients are greatly influenced by the

manner in which the doctor presents treatment options. Orentlicher reviews
studies indicating that patients are more likely to choose surgery when the
probability of survival is presented rather than the probability of death,'"*' and

138. See JAY Katz, Silent World of Doctor and Patient (1984). Katz presents the
classic critique of informed consent as a legal requirement and as a professional practice. In his

analysis ofthe legal requirement of informed consent, Katz calls informed consent a "fairy tale" and

delineates why informed consent is unnatural to physicians and how the law's enforcement of the

requirement is ineffective and half-hearted at best.

1 39. David Orentlicher, The Illusion ofPatient Choice in End-of-Life Decisions, 267 JAMA
2101,2101 (1992) (emphasis added) (citations omitted).

140. Id. (discussing Marion Danis et al., A Prospective Study ofAdvance Directivesfor Life-
Sustaining Care, 324 NEW Eng. J. MED. 882 (1991)).

141. /d/. at 2102 (discussing T. M. Marteau, Framing ofInformation: Its Influence Upon

Decisions ofDoctors and Patients, 28 BR. J. SOC. PSYCH. 89 (1989)).



1998] END-OF-LIFE DECISION MAKING 35

that patients are more likely to choose treatment when the treatment is presented
positively rather than negatively.'"*^

Orentlicher also reviews studies that indicate that physicians "are more
inclined to talk with patients who are most like them"'"*^ and that physicians give
both more time and more explanations of the course oftreatment to patients who
seem more intelligent and better educated.'"*^ Orentlicher concludes: "Ironically,

physicians may become most aware ofthe preferences of patients who share their
values. . . . Physicians may be less aware of the preferences of patients whose
values diverge sharply from those of their physicians."'"*^

More recent research remains consistent with the research relied upon by
Orentlicher in his 1992 article.'"*^ In particular, the results ofthe SUPPORTstudy
of in-hospital end-of-life care, discussed later in this article, confirms

Orentlicher' s conclusions.'"*^

Empirical research on the medical neglect of pain and on non-compliance

with patients' treatment choices at the end of life contradicts the Ninth Circuit's

reliance on physicians as a safeguard in the administration of legalized physician-

assisted suicide. The data actually support the proposition that physicians will
not act as a guard against suicides undertaken because of neglected pain or

without the complete understanding and consent of the patient.

Steven Miles calls for more study of the psychology ofthe physician-patient

relationship in relation to physician-assisted suicide. He notes:

There is ample evidence that physicians often have difficulty

responding therapeutically to chronically ill or dying patients. They

often underdiagnose and undertreat pain or depression, either of which

may be part ofwhy a patient may choose to die. Chronically ill or dying
patients have told how their physicians emotionally or physically
withdrew from them. ...

The intractability of these findings, despite years of calls for more
sensitivity to the needs of dying persons and for better use of therapeutic

information, implicates deep causes in the genesis of such mistreatment.

... In this view, a physician tends to withdraw from dying patients,

minimize pain, overlook clinical depression, and prematurely believe

142. Id. (discussing T.R. Malloy et al., How Interventions Are Described Affects Patients

'

Decisions About Life-Sustaining Treatment, in AMERICAN GERIATRIC Society/American

Federation for Aging Research Annual Scientific Meeting, 1991 Program (Abstract A2)).

143. Mat 2102.
144. Id

145. Id

146. See, e.g., Lawrence J. Schneiderman et al., Do Physician 's Own Preferencesfor Life-
Sustaining Treatment Influence Their Perceptions ofPatients ' Preferences? A Second Look, 6
Cambridge Q. Healthcare Ethics 131 (1997); Steven H. Miles, Physicians and Their Patients

'

Suicides, 271 JAMA 1786 (1994).
147. See discussion infra Part II.A.



36 INDIANA LAW REVIEW [Vol. 31:13

patients have chosen to die as a result of the physician's own dread of
death or disability or belief that a patient is as demoralized by his or her
condition as the physician is. If this critique is correct, physicians are

not bad, just human.'"**

Descriptions of current medical practices do not necessarily predict future

practices. The patterns described by Orentlicher and SUPPORTand others might
be remedied in the future. The Ninth Circuit relies on current medical training
and ethics as a source of constraint for physicians in assisting death, and the

court anticipates a future in which medical training and ethics will continue to

operate in that fashion and will continue to emphasize constraints or prohibitions

on intervening to cause death. Opponents to legalization argue that legalization

itself will change medical ethics and medical education in the most significant

way.'"*^ Perhaps these contrasting visions of the future could be empirically

tested, although problems in structuring such a study seem insurmountable.

Even if such an experiment could be structured, it is unlikely that the data
would resolve the question of whether the legalization of physician-assisted
suicide has a positive or negative impact on physician behavior and medical

ethics. Proponents of legalization argue that physicians who now cannot aid a
patient in suicide without committing a crime, abandon their patients and their

own ethical duty to relieve suffering.'^^ In this view, the ethical physician is one
who is willing to care for the patient even to the point of ending that patient's
life, and a society that respects an individual's choice to end his life is a better

society. Opponents argue that the physician's ethical duty to the patient requires

that the physician be fully committed to the well-being of the patient, including

aggressively treating pain even to the point of the patient's unconsciousness, but

that the doctor's duty requires refusing the patient's request for death. From this
perspective, a society that allows physicians to kill patients upon request is a

society that has abandoned caring for the ill, the dying and dependent.'^'

Empirical research could measure physicians' behavior pre- and post-

legalization, but it cannot answer whether the behavior is more or less ethical
than before.

The Ninth Circuit's confidence in physician practices in respecting their

148. Miles, supra note 146, at 1788-89.

149. Reply Brief for Petitioners, Washington v. Glucksberg, 1 17 S. Ct. 2258 (1997) (No. 96-

1 10); Brief of Family Research Council as Amicus Curiae in Support of Petitioners, Washington

V. Glucksberg, 1 17 S. Ct. 2258 (1997) (Nos. 96-1 10, 95-1858).

1 50. Brief of the American Medical Student Association and a Coalition of Distinguished

Medical Professionals as Amici Curiae in Support of Respondents, Vacco v. Quill 1 17 S. Ct. 2293

(1997) (Nos. 95-1858, 96-110); Brief of Amicus Curiae Bioethicists Supporting Respondents,

Vacco V. Quill 1 17 S. Ct. 2293 (1997) (Nos. 95-1858, 96-1 10).

151. Brief of the American Medical Association et al. as Amicus Curiae in Support of

Petitioners, Vacco v. Quill, 1 17 S. Ct. 2293 (1997) (No. 95-1858); Brief of the International Anti-

Euthanasia Task Force in Support of Petitioners, Vacco v. Quill, 1 17 S. Ct. 2293 (1997) (Nos. 95-

1858,96-110).



1998] END-OF-LIFE DECISION MAKING 37

patients' wishes concerning medical treatment and in treating patients in pain is

clearly misplaced. Part of the Ninth Circuit's argument in favor of a right to

physician-assisted death, however, is that withdrawal of life-sustaining treatment

under the current legal framework itself presents risks of abuse. '^^ This

observation could logically lead to restrictions on withdrawal of life-sustaining

treatment just as much as it could lead to acceptance of the use of lethal drugs.
The argument that there is no difference between the two can work both ways.

C At-Risk Populations
Opponents to legalization of assisted suicide argue that legalization would

place certain vulnerable populations at particular risk of involuntary, coerced, or

simply desperate decisions for termination of life.'^^ The groups that are often
considered particularly vulnerable in the U.S. health care system include the

poor, minorities, those disadvantaged by lack of access to adequate health care,

the disabled and the elderly.'^"* It is common practice for courts to identify
particular groups as vulnerable and to construct legal rules designed to

compensate for that vulnerability.

The Ninth Circuit reviewed arguments that these vulnerable populations will

be at risk. However, the court of appeals rejected arguments that most of the

vulnerable populations identified will be negatively affected by a change in the

legal status of physician-assisted suicide. The court makes several different

arguments on this point.

The court first addresses the argument that persons who do not have access
to adequate health care will be threatened by the legalization of physician-

assisted suicide:

The argument that disadvantaged persons will receive more medical
services than the remainder of the population in one, and only one

area—assisted suicide—is ludicrous on its face. So, too, is the argument
that the poor and the minorities will rush to volunteer for physician-

assisted suicide because of their inability to secure adequate medical

treatment.
^^^

The Ninth Circuit is particularly harsh in its response to such arguments. It
supports its conclusion with the experience of access to abortions, an instance

where similar arguments about vulnerable populations have been made.'^^ The
court notes that public funding for abortions is very limited and that other factors

152. Compassion in Dying v. Washington, 79 F.3d 790, 817 (9th Cir. 1996).

153. Reply Brief for Petitioners Vacco and Pataki, Vacco v. Quill, 1 17 S. Ct. 2293 (1997)

(No. 95-1858); Brief ofthe American Geriatrics Society as Amicus Curiae Urging Reversal of the

Judgments Below, Vacco v. Quill, 1 17 S. Ct. 2293 (1997) (Nos. 95-1858, 96-1 10).

154. Compassion in Dying, 19 VMdX ^15.
155. Id.

156. Id.



38 INDIANA LAW REVIEW [Vol. 31:13

converge to limit access even where there is an ability to pay.'^^

One could argue that the case of abortion services is not strictly analogous
to this case. Abortion services are more costly and require more institutional
collaboration, in terms of space and equipment, than would a prescription for
lethal medication. Although abortion might not be a totally effective analogy,

there may be other analogous situations.
Whether or not abortion is analogous to assisted suicide, a focus on the

anticipated utilization rate reduces the argument over whether legalization

specially endangers those who are currently denied health care and the poor and
minorities to an empirically verifiable question. A study could be constructed to
monitor the provision of assisted death services in a way that would test the
court's assumption about excess utilization by such groups. Structuring the study

would be difficult unless the use of the service could be both tightly controlled
and accurately recorded, but it still might be feasible.

'^^

Arguments based on inadequacies in the health care system include a further

assertion that legalization of assisted suicide may undercut already weak public
support for adequate health care, particularly for conditions that can trigger

substantial costs with little or no hope for a return to health or substantially

improved functioning. This concern reaches care for the dying and long-term

care for the seriously debilitated elderly, among other situations.'^^ This view
argues that if the sometimes drawn-out "natural dying" process is viewed as a

matter of personal choice, those who choose natural dying will have to bear the
cost of that choice. The proposed regulation of assisted suicide in the Oregon
initiative of 1994 attempts to respond to this problem and prohibits

[t]he sale, procurement, or issuance of any life, health, or accident

insurance or annuity policy or the rate charged for any policy shall not

be conditioned upon or affected by the making or rescinding of a request

by a person, for medication to end his or her life in a humane and
dignified manner.

'^°

Such regulation, however, cannot control the general level of insurance coverage

or government payment for health care services for the dying, disabled or very
dependent elderly.

Dissipation of public and private support for health care and supportive care

for the dying and the debilitated elderly would be quite hard to measure. Such
a change may operate glacially, and not be detectable for many years. This
concern over withdrawal of financial support—^whether public or private—is
speculative, but so are assertions that legalization of assisted suicide will not

157. Id.

1 58. See infra text accompanying notes 164-65 (discussing various approaches to fairness).

1 59. See, e.g., Susan Kastner, The Battle Between Generations Many ofToday 's Sandwich
Generation, the Newly Naked Middle Class, Have Been Forced Back to Frequently Bitter

Interdependence. Young Against Old, Old Against Young, Families Cloistered Together, and at

Odds, Toronto Star, April 22, 1996, at A15.

160. Or. Rev. Stat. § 127.875 (1996).



1998] END-OF-LIFE DECISION MAKING 39

have such an impact.

The Ninth Circuit also addresses concerns about discrimination against the
disabled. At one point, the court opined that such discrimination is unlikely

because "[o]rganizations representing the physically impaired are sufficiently

active politically and sufficiently vigilant that they w^ould soon put a halt to any

effort to employ assisted suicide in a manner that affected their clients

unfairly
y^^^ The court did not describe what it meant by unfair treatment.

The situation ofthe disabled quite clearly raises the issue of what constitutes

discrimination against any population. Is it discriminatory to legalize assisted

suicide by disabled persons; or is it discriminatory to "protect" such individuals

by excluding them from choosing certain options any other individual may
choose? The Ninth Circuit in Compassion in Dying responded in this fashion:

[sjeriously impaired individuals will, along with non-impaired

individuals, be the beneficiaries ofthe liberty interest asserted here—and
. . . if they are not afforded the option to control their own fate, they like
many others will be compelled, against their will, to endure unusual and
protracted suffering.

^^^

The court's perception of assisted suicide as beneficial is a value judgment and

is unlikely to be resolved empirically.

The Ninth Circuit offered fairness as a basis on which to resolve the conflict

between risks to vulnerable populations generally and limitations on individual

liberty: "The resolution that would be best for all, of course, would be to ensure

that the practice of assisted suicide is conducted fairly and well, and that

adequate safeguards sufficient to avoid the feared abuses are adopted and

enforced."^^^ In this argument, legalization would allow regulation of assisted

suicide and activate safeguards to prevent abuses.

The aspiration that the system be administered "fairly and well" appears to

be a standard that can be performance-monitored. One common and well
accepted method for monitoring the fairness of any allocation system is to

examine patterns in utilization. In recent years, for example, many studies on
health care delivery have produced statistics on the provision of certain medical

interventions across populations.'^'^

Empirical studies on the incidence of assisted suicide across identified

populations must operate on implied values of fairness: for example, that a

161

.

Compassion in Dying, 79 F.3d at 825 (emphasis added).

162. Id.

163. Mat 825.
164. See, e.g., M.B. Hamel et al., Seriously III Hospitalized Adults: Do We Spend Less on

Older Patients?, 44 J. AM. GERIATRICS SOC. 1043 (1996); Scott Burris, Dental Discrimination

Against the HIV-infected: Empirical Data, Law and Public Policy, 13 YALE J. ON REG. 1 (1996);
Carol Jonann Bess, Gender Bias in Health Care: A Life or Death Issuefor Women with Coronary

Heart Disease, 6 HASTINGS WOMEN'S L.J. 41 (1995); David R. Williams & Chiguita Collins, U.S.
Socioeconomic and Racial Differences in Health: Patterns and Explanations, 21 ANN. REV. Soc.

349(1995).



40 INDIANA LAW REVIEW [Vol. 31:13

distribution that is determined solely by race or gender would be "fair" or
"unfair." Empirical data standing alone cannot resolve the question of whether

the system for assistance in death is operating "fairly." The resolution of
whether assisted death is administered fairly requires that the content of the

notion of fairness be made more specific. If fairness is taken in its most common
meaning of "treating like cases alike," then the criteria for relevant similarities
and differences need to be identified to provide the Ninth Circuit's notion of
fairness with some content.

It is not clear from the Ninth Circuit's opinion, for example, how data that
indicate a disproportionate representation of "seriously impaired individuals"

among those choosing assisted suicide should be assessed. Would the data be
interpreted as showing that assisted death was being administered fairly because
it was reaching the people who could significantly benefit from the service, or
would the data provide prima facie evidence of unfairness because of concerns

that the disabled were not receiving adequate health care? The court's
observation that "seriously impaired individuals" will be among the beneficiaries
of legalization indicates that the court would not view disproportionate
representation ofdisabled persons among those receiving medical services to end
their lives as a problem.

The influence ofeconomic factors in the administration of physician-assisted
suicide might also be studied. Hypothetically, data could indicate either that

patients receiving physician aid in death were primarily from higher income

groups or that uninsured patients were disproportionately represented. If assisted

suicide patients were often uninsured, would that indicate unfairness in the

system? If only higher income patients were using this service, would it indicate

that there were financial or non-financial barriers, such as required consultations

and psychiatric assessments, that unfairly excluded lower-income uninsured

patients? The Ninth Circuit opinion directly addressed the issue of financial
considerations. In response to arguments that the terminally ill might choose

assisted suicide because of financial pressures from "astronomical medical bills"

that "consume the life savings they planned to leave for their families" or that
"worse yet, burden their families with debts they may never be able to satisfy,"
the court stated that "[w]e are reluctant to say that, in a society in which the costs

of protracted health care can be so exorbitant, it is improper for competent,

terminally ill adults to take the economic welfare oftheir families and loved ones

into consideration."'^^ It is fair to say that the Ninth Circuit would not consider

a disproportionate representation of persons of lower economic status an

indication ofunfairness in the administration of physician-assisted death. In fact,

the court, in rejecting fears that the poor and minorities would be at risk ofabuse,

stated that there "is far more reason to raise the opposite concern: . . . that the
poor and the minorities . . . will not be afforded a fair opportunity to obtain . . .

the assistance that would allow them to end their lives with a measure of
dignity."'^^

165. Compassion in Dying, 79 F.3d at 826.

166. Mat 825.



1 998] END-OF-LIFE DECISION MAKING 4

1

The court did not address race-based patterns separately from income-based
patterns, although the two can operate separately on access to health care.'^^ Nor
did the court separately address gender, although gender differences in health

care have been discussed in the context of life-sustaining treatment decisions and

physician-assisted suicide/^^

II. The Study to Understand Prognoses and Preferences for
Outcomes and Risks of Treatments (SUPPORT)

If money were no object, could we design an empirical study that would
determine once and for all whether we should maintain our allegiance to the ideal
of patient choice in medical treatment decision making at the end of life? And
if we could design such a study, would it resolve the issue of whether patient
choice will drive physician-assisted death or whether physicians will drive

physician-assisted death?

The Study to Understand Prognoses and Preferences for Outcomes and Risks
of Treatments, '^^ commonly known as SUPPORT, was a very large and very
ambitious empirical study ofthe principles of bioethics in practice and certainly

the most massive to date. SUPPORT, funded by the Robert Wood Johnson
Foundation for $28 million,'^° studied the care of 9,105 patients admitted to five

major medical centers in five states from June 1989 through January 1994. Eight

co-principal investigators, with the advice of a ten-member national advisory
committee, directed the efforts of over 150 professional staff engaged in the

project.

Each ofthe over 9,000 patients was admitted to a participating hospital in the
advanced stages of one of nine life-threatening or terminal illnesses. SUPPORT
studied the care these patients received while in the hospital with a particular

focus on resource utilization and on communication of and compliance with

patient preferences. In their principal report on the project,'^' the investigators
clearly view the importance of the study in relation to the basic principals of
bioethics, especially as they have come to be understood by clinicians and
institutionalized by the law in end-of-life care. The investigators note that
"professional organizations, the judiciary, consumer organizations, and a

167. See, e.g.. Council on Ethical and Judicial Affairs, American Medical Association, Black-

White Disparities in Health Care, 263 JAMA 2344 (1990); Vemellia R. Randall, Racist Health
Care: Reforming an Unjust Health Care System to Meet the Needs of African-Americans, 3

Health Matrix 127(1993).

1 68. Steven H. Miles & Allison August, Courts, Gender and "The Right to Die, " 1 8 L. MED.
& Health Care 85 (1990); Nancy J. Osgood & Susan A. Eisenhandler, Gender and Assisted and
Acquiescent Suicide: A Suicidologist's Perspective, 9 ISSUES L. & Med. 361 (1994).

1 69. SUPPORT, supra note 25, at 1 59 1

.

170. Steven A. Schroeder, Robert Wood Johnson Foundation, President's Message, On
Dying in America, <http://www.rwjf.org/library/annual/pressrl.htm>.

171

.

The study produced many articles, but the principal report was published in the Journal
of the American Medical Association. See SUPPORT, supra note 25.



42 INDIANA LAW REVIEW [Vol. 31:13

president's commission have all advocated more emphasis on realistically
forecasting outcomes of life-sustaining treatment and on improved

communication between physician and patient."^^^ The investigators view
SUPPORT as an empirical test of the operation of informed consent and
compliance with patients' preferences. In addition, the study associated patient

choice with the issue of resource consumption in hospital care for patients near

the end of life.
^'^

The first years of SUPPORT, now known as Phase I, were conducted from
June 1989 to June 1991. According to the investigators, Phase I revealed

"shortfalls in patient-physician communication."^^"^ For example, of the 4,301

patients enrolled in the protocol in those years, 31% preferred that cardio-
pulmonary resuscitation be withheld in the case of cardiac arrest, but fewer than

half of the physicians treating these particular patients knew of their patients'
preference to forego resuscitation.'^^ For only 47% of all ofthe patients enrolled
in Phase I could their physicians accurately report the patient's choice whether

to accept or refuse resuscitation. Somewhat fewer than half of those patients
who preferred not to undergo CPR had a written order (a "do-not-resuscitate
order" or "DNR") which would allow CPR to be foregone should they arrest. '^^

Because most hospitals require that CPR be performed unless there is a written
order otherwise, the patients without such an order may have undergone
unwanted resuscitative interventions if they had suffered an arrest.'^^

Pain was undertreated. Twenty-two percent of the patients interviewed
reported suffering moderate to severe pain at least half the time in the hospital.

Surrogates or family members interviewed after a patient had died reported that
half of all conscious patients who had died in the hospital experienced moderate
to severe pain at least half of the time over the last three days of

hospitalization.'^^

The investigators reported that there was a substantial variation in these
results among physician specialty groups and among the five medical centers.
For example, the percentage of patients reporting moderate to severe pain ranged

from a low of 12% in one institution to a high of32% in another. Agreement on
reports of DNR preferences varied by specialty from 8% for cardiologists and
their patients to 24% for oncologists and their patients. '^^

The study was monitored throughout its course. Early findings revealed
these substantial problems in physician-patient communication, treatment

decisions and outcomes. When doctors at the participating hospitals indicated

172. Mat 1591.
1 73. Id The study used total hospital resource consumption per case and the number of days

spent in intensive care when the patient was on ventilator support or in a coma as indicators. Id.

174. Mat 1592.
175. Mat 1594.
176. Id

177. Id

178. Id

179. Id



1998] END-OF-LIFE DECISION MAKING 43

an interest in trying to change the situation, the study was altered in an attempt

to change these observed patterns at the participating hospitals. Meetings were

held with the doctors and other participants to develop a plan for improving

physician-patient and family communication and for bringing treatment

decisions, such as the decision concerning resuscitation, in alignment with

patient preferences and probable outcomes. The result of these meetings became
Phase II oiSUPPORT}^''

According to investigators, physicians participating in Phase I asserted that

communication would improve if they could have access to more reliable and
timely information and if project staff "would make it more efficient to have
conversations."'^' In response to these ideas, the project designed an intervention

to respond to the doctors' expressed needs. In Phase II of the project, patients

were divided into a control group and an intervention group. '^^ Patients and

physicians in the intervention group received services from a SUPPORT nursQ
specifically focusing on providing information to the patient and family

concerning treatment and prognosis and eliciting discussions with them
concerning the patient's or surrogate's choices for treatment.'^^ The patient's
physician had to approve before the SUPPORT nurse could work with any
particular patient.

'^"^

SUPPORTnursQS worked with nearly all ofthe 2,652 patients enrolled in the
intervention group of Phase II through a randomized process following hospital

admission. Only 133 patients enrolled in the intervention group did not receive

the /S'L^PPOi?r intervention, and ofthese, 75 had died or were discharged on the

day they were placed in the study. '^^ The SUPPORTnursQ intervened as seemed
advisable in each case. The SUPPORT nuvsQ talked directly with the patient or
family about the patient's prognosis in 84% of the cases; about pain in 77%;
about likely outcomes of resuscitation in 63%); and about written advance

directives in 73% of the cases.'^^
ASUPPORTnursQ discussed the patient's choices and understanding oftheir

medical condition with the patient's doctor in "virtually all cases." At least one
written report ofthe patient's or surrogate's preferences was provided in 78%) of
the cases. '^^ SUPPORT nurses also engaged in "time-consuming discussions,
arranged meetings, provided information, supplied forms, and did anything else

to encourage the patient and family to engage in an informed and collaborative

decision-making process with a well-informed physician."'^^

SUPPORTalso provided doctors in Phase II with information on the patient's

180. /fl?. at 1592.

181. Id.

182. Id. at \593.

183. Id at 1592.

184. Id

185. /fl?. at 1594.

186. Id

187. Id

188. Id



44 INDIANA LAW REVIEW [Vol. 31:13

prognosis. SUPPORTdevQloped a prognostic model for predicting survival rates
and outcomes for the types of patients involved in the study. For 94% of the
patients in Phase II, physicians received at least one written prognosis evaluation,

and this written evaluation was placed in the patient's medical record in 80% of
the cases.

'^^

A. The Results ofS\J??ORT

The investigators reported on the results of Phase II: "In phase II of
SUPPORT, improved information, enhanced conversation, and an explicit effort
to encourage use of outcome data and preferences in decision making were
completely ineffectual, despite the fact that the study had enough power to detect

small effects."'^° In describing the detail that supported their conclusion,

SUPPORT mwQstigators reported that only 34% of the physicians acknowledged
having received a report of patient preferences, although a written report was
provided in 78% of the cases. '^' Fifty-nine percent of the doctors acknowledged
having received the prognosis report, even though a written report was given to
the doctor in 94% of the cases and had been placed in the patient's medical
record in 80% of the cases. '^^ Only 15% of the participating doctors reported
having discussed the information on prognosis and preferences with their

patients. '^^ The prevalence and timing of written DNR orders was the same for
the control and the intervention groups. The investigators reported that there was

"a small association" of the intervention with improvement in patient-physician

agreement on the patient's desires concerning resuscitation.'^'* There was no
change in the use of hospital resources or in the length of stay in the intensive

care unit. There was an increase in reported untreated pain in the interviews
conducted with patients themselves. *^^ The intervention did not change the
unadjusted proportion of patients or surrogates who reported having a discussion
about CPR (37% of the control group and 40% of the intervention group), with
41% of those reporting that CPR had not been discussed stating that they would
have liked to discuss the decision.

'^^

B. Reactions to SU??OKT

The President of the Robert Wood Johnson Foundation said that "the
investigators were stunned" with the results of Phase 11.'^^ He reported that he
was not surprised with the results, however, because of his own experiences as

189. Id at 1595.

190. Id. at \ 596.

191. /^. at 1595.

192. Id

193. Id

194. /^. at 1596.

195. Id

196. Id

1 97. Schroeder, supra note 1 70.



1 998] END-OF-LIFE DECISION MAKING 45

a physician and because of the "horror stories" told him by relatives and friends.
The investigators themselves stated: "We are left with a troubling situation. The
picture we describe of the care of seriously ill or dying persons is not attractive."
They also described the ideal to which they compared the Phase II results: "One
would certainly prefer to envision that, when confronted with life-threatening
illness, the patient and family would be included in discussions, realistic
estimates ofoutcome would be valued, pain would be treated, and dying would
not be prolonged."*^^ This was the investigators' vision of good care for the
dying. But the investigators observed that "most patients and families indicated

they were satisfied [with the care], no matter what happened to them."'^^

The investigators offered their own advice for the appropriate response to the
results ofSUPPORT. They conclude their report by arguing that improved care
for the dying and improved communication require "reexamination of our

individual and collective commitment to these goals, more creative efforts at
shaping the treatment process, and perhaps, more proactive and forceful attempts
at change."2°'

The Robert Wood Johnson Foundation, which funded SUPPORT, took a
"proactive" and "forceful" action in reaction to the results ofSUPPORT The
Foundation established a national campaign to improve care of the dying. This

"Last Acts" campaign is a nationwide multi-million-dollar effort to improve care

of the dying through funded research, demonstration projects, consensus

conferences, task forces, and media events. The Foundation also funded the first

round of expert commentary on SUPPORT, published in the Hastings Center
Report to coincide with the release of the results of Phase 11.^°'

C. Does S\JF?ORT Answer Our Questions?

SUPPORThas revealed a gap between behavior and normative expectations.
Current normative expectations require that physicians talk with their patients or

the patients' surrogates, about prognosis, treatment options and choices. The
same norms expect that patients will participate in such discussions and perhaps
will want to have such discussions. Norms requiring discussion include an
expectation that physicians will account for, if not follow, the patients' treatment

preferences. These are not new ethical and legal standards. They have provided
the dominant ethical and legal framework for life-sustaining treatment decisions

for the past two decades and for informed consent to treatment since at least the

mid-1950s, if not earlier.^°2

1 98. SUPPORT, supra note 25, at 1 597.

199. Mat 1596.
200. /^. at 1597.

20 1

.

Schroeder, supra note 1 70.

202. Justice Cardozo's opinion in Schloendorffv. Society ofNew York Hospital, 105 N.E.
92 (1914), is often cited as the root of the right to refuse treatment. In that opinion, Cardozo stated:

"Every human being of adult years and sound mind has a right to determine what shall be done with

his own body." Id. at 129. For one history and a critique of legal enforcement of informed consent.



46 INDIANA LAW REVIEW [Vol. 31:13

Does SUPPORTtcW us that patient choice concerning life-sustaining medical
treatment is an ideal that should be abandoned? SUPPORTcertainly reveals that
the intervention used in Phase II to increase conversation and information did not

improve the incidence of patient-physician communication substantially; that it

only minimally improved the accuracy of physician knowledge of patient
preferences; that it did not alter physician decisions concerning writing a DNR
order or discharging a patient from an ICU; that it did not improve treatment for

pain; and that the great majority of both physicians and surviving family

members reported feeling satisfied that things went well. But many of those
evaluating the SUPPORTresults charge this failure to defects in the intervention
itself: the intervention put an intermediary between the physician and the

patient;^°^ the intervention used nurses instead of physicians as the stimulus for

change;^°^ the intervention did not try to alter the organizational environment and

incentives for maintaining current behaviors;^^^ the Project's prognosis model
produced results no more accurate than physician judgment and so could not
change decisions that would have been made by physicians without the model;^^
and so on.

Some argue that another intervention designed differently could succeed^°^
and that the lesson ofSUPPORT is that we must redouble our efforts to assure
that behavior conforms to our ideals. In this view, SUPPORT is not the death
knell for patient autonomy and informed consent in decisions concerning life-

sustaining treatment; it is argued that more education and more commitment can
change the patterns revealed by SUPPORT.

The more radical responses to SUPPORThawc argued that SUPPORT is not
groundbreaking. SUPPORTioins a great body of evidence and analysis that
proves that informed consent, conversation and patient autonomy is ill-suited to
decision making regarding medical interventions. Physicians simply will not talk

with their patients, perhaps especially in terminal care, and will not yield control

to the patients. Doctors may not be the only ones resistant to planning for
death.^°^ Some argue that it is time to abandon the myth of individual patient

see Katz, supra note 138. See also Schuck, supra note 4, at 900 (observing that the legal doctrine

did not emerge in "relatively robust form" until 1957).

203. Ellen H. Moskowitz & James L. Nelson, The Best Laid Plans, 25 Hastings Ctr. Rep.
S3 (1995).

204. Id

205. Mildred Z. Solomon, The Enormity ofthe Task: SUPPORTand Changing Practice, 25
Hastings Ctr. Rep. S28 (1995).

206. On the reliability of the prognosis report as compared with the individual physicians'
judgments, see Joanne Lynn et al., Accurate Prognostications of Death: Opportunities and

Challenges for Clinicians Caring for Patients at the End of Life, 163 W.J. MED. 250 (1995);
Bernard Lo, Improving Care Near the End ofLife: Why Is It So Hard?, 21A JAMA 1 634 ( 1 995).

207. Bernard Lo, End-of-Life Care After Termination ofSUPPORT, 25 HASTINGS Ctr. Rep.

S6 (1995); Solomon, supra note 205.

208. Ellen Goodman, The Conspiracy ofSilence About Death is Too Strong, Baltimore

Sun, Dec. 1, 1995, at 27A (reporting on family and patient resistance); Special Report: Critical



1 998] END-OF-LIFE DECISION MAKING 47

autonomy, or consumer choice, and redesign the relationship between physician

and patient along different and yet to be announced lines.^°^

Conclusion

While researchers design the "next SUPPORT^ to further test the workability
of patient choice in end-of-life care, legislatures and courts are being asked to

regulate the area. Empirical research can certainly aid in this effort, but will not

resolve the ultimate issues. Empirical research does not have the answers. It is

likely that the studies will never be well-enough designed or the results

conclusive enough to predict the direction and impact of change.

Furthermore, the tension between what is now and what should be cannot be
resolved by the data. The best that can be hoped for is that assumptions are
recognized for what they are rather than accepted with the confidence of fact and

where empirical data do indicate patterns of behavior, those patterns are
recognized rather than denied.

What we do not know we can make up, but we should all realize when that
is being done. What we do know, as little as that may be, we cannot ignore. It
would be a great loss, if advocacy for legalization of physician aid in dying or

assisted suicide triggered the collapse of the social pact that allows us to stop

aggressive treatment of terminally ill patients, vegetative patients, or patients for

whom treatment causes pain without benefit. But it also seems foolhardy to
legalize assisted death on the basis of a fantasy of physician commitment to

patient choice in health care and to the provision of adequate pain relief.

Timesfor Critical Care: What About "Quality ofDeath, " MED. OUTCOMES & GUIDELINES ALERT,
Nov. 30, 1995 (describing need for changing institutional culture rather than physician attitudes).

209. See, e.g. , Eric G. Anderson, Some New Lessons to Consider About Life—and Death, Am.
Med. News, Aug. 5, 1996, at 35.