Leslie P. Francis 

Leslie P. Francis is Alfred C. Emery Professor of Law; Professor and Chair, 
Department of Philosophy; and a member of the Division of Medical Ethics 
in the Department of Internal Medicine at the University of Utah. She 
received her Ph.D. in Philosophy from the University of Michigan in 1974 and 
her J.D. from the University of Utah in 1981. She was a law clerk to Judge 
Abner Mikva of the United States Court of Appeals for the District of 
Columbia Circuit in 1981-82. 

Professor Francis specializes in Ethics, Bioethics, Philosophy of Law, Health 
Law, and Disability Law. She is the author of a number of articles on issues 
in philosophy of law, health care, and professional ethics. Her most recent 
books are LAND WARS: THE POLITICS OF PROPERTY AND COMMUNITY (with 
John G. Francis, Lynne Rienner Publishers 2003), and SEXUAL HARASSMENT 
AS AN ETHICAL IsSUE IN ACADEMIC LIFE (Roman and Littlefield 2001). She 
also has edited (with Anita Silvers) AMERICANS WITH DISABILlTIES: 
IMPUCATIONS OF Tiffi LAW FOR INDIVIDUALS AND INSTITUTIONS (Routledge 
2000). She is currently a member of the American Law Institute (elected 
1986), Chair of the Program Committee for the Pacific Division of the AP A, 
and Immediate Past President of AMINTAPHll.... In the spring of 2000, she 
was awarded the University's Rosenblatt Prize for overall excellence in 
research, teaching, and academic service. 





LEGITIMATE ExPECTATIONS, UNREASONABLE BELIEFS, 
AND LEGALLY MANDATED COVERAGE OF 

EXPERIMENTAL THERAPY 

Leslie Pickering Francis* 

INTRODUCTION 

Photographs of patients seeking contributions for expensive bone 
marrow transplants are an everyday image on supermarket checkout stands. 
Benefit concerts, newspaper stories, and community fundraisers pitch in to 
help patients who cannot otherwise afford expensive medical interventions. 
Patients with multiple myeloma lined up to testify before the Medicare 
Coverage Advisory Committee about the lifesaving possibilities of "their'' 
bone marrow transplants.' Many of the patients who are the subjects of these 
appeals have health insurance, but their insurance has denied coverage 
because the care has been judged to be experimental or not reasonably 
medically necessary. Behind these appeals for funding lies the judgment that 
these patients are not receiving care to which they are entitled-at least, care 
to which they are sufficiently entitled to deserve community support for their 
fundraising efforts. 

Are these patients justified in their claims? Should health insurance 
routinely cover the costs of experimental therapy? Do patients have a moral 
claim that their legitimate expectations have been disappointed when the care 
is not covered? Patients and their advocates understandably urge coverage, 
especially when the therapy is expensive and lives are apparently on the line. 
In contrast, understandably, insurance plans would prefer to limit coverage for 
expensive interventions of marginal value. Ideally, all are interested in ensur­
ing that patients receive proper care while also reducing the likelihood of 
ineffective or dangerous interventions. In the midst of this debate, a number 
of states have recently enacted coverage mandates for off-label drug uses and 
for patient participation in clinical trials (see Part IV). One policy goal of at 
least some of these statutes is to ensure that patients receive therapeutically 
beneficial care. Alabama, New Jersey, and Tennessee, for example, mandate 
coverage for patients receiving prescriptions for certain off-label drug uses; 
their stated legislative purpose is that the patients' "expectation that their 

* Leslie Pickering Francis, Ph.D., University of Michigan, 1974; J.D., University of 
Utah, 1981; Alfred C. Emery Professor of Law and Professor and Chair, Department of 
Philosophy, University of Utah. 

1. Drugs, Biologics, & Therapeutics Advisory Panel, Medicare Coverage Advisory 
Comm., Minutes of the June 15, 1999 Meeting, available at http://www.cms.gov/mcac/Sb 1-
d5.asp? origin=globalsearch&page=/mcd/viewmcac.asp?id=lO. 



216 INDIANA HEALTH LAW REVIEW [Vol. 1:213 

health care costs will be paid by their insurance policies is not disappointed. "2 

This statutory policy apparently assumes that patients' legitimate expectations 
for therapeutic care extend to the covered off-label uses. An additional 
statutory justification for the coverage mandates is the desire to encourage the 
development of evidence-based medical knowledge about new therapies. 
Further justifications are drawn from distributive justice concerns-all patient 
populations should be included in the development of new medical knowl­
edge, and promising new therapies should be available on an equitable basis. 

This Article explores whether "legitimate expectations" for treatment 
are a persuasive moral reason to mandate insurance coverage of experimental 
therapy. It argues that legitimate expectations for coverage of therapeutic care 
extend only to limited cases of experimental interventions. Some of the statu­
tory mandates exceed this justification in significant ways. Reasons based in 
justice and in evidence-based medical knowledge support some mandated 
coverage, but this logic does not rest primarily on the patient's own entitle­
ment to care. Moreover, the encouragement of patients' expectations that ex­
perimental care will be funded may continue to induce problematic litigation 
by patients seeking the coverage. A principal example is coverage of high­
dose chemotherapy with stem cell support for patients with breast cancer, 
which is a politically and emotionally charged issue but is by no means 
unique. 

I. THE REASONABLENESS OF ExPECTATIONS AND THE 
LEGITIMACY OF BELIEFS 

This essay is part of a larger project exploring the moral import of 
expectations. The goal of the larger project is to explore when, if ever, and 
why the fact that people have formed expectations-beliefs about the future 
that they treat with an attitude of reliance-matters morally to how they ought 
to be treated. Respecting expectations is an aspect of treating people as 
choosers and planners of their lives; it is part of what we must do if we are to 
respect autonomy in practice. But not all expectations matter, and some that 
do matter bear only limited moral weight. One critical aspect of whether and 
with what weight expectations make a difference to how people ought to be 
treated is whether the expectations are based in reasonable beliefs. This 
aspect of expectations-the reasonableness of the beliefs on which they are 
based and their resulting moral significance-is the subject of this Article. 

In enacting these statutes, the legislatures of Alabama, New Jersey, and 
Tennessee stated explicitly that their goal was to protect the legitimate 
expectations of patients. These expectations, as the legislatures understood 
them, rested on two related sets of beliefs. First, patients-and the legislators 
who acted to support them-believed the off-label therapies at issue were 

2. ALA. CODE§ 27-1-10.1 (2003);N.J.STAT.ANN. § 26:1A-36.9(b) (West2003); TENN. 
CODE ANN.§ 56-7-2352 (2003). 



2004] MANDATED COVERAGE OF ExPERIMENTAL THERAPY 217 

likely to be of significant medical benefit. 3 Second, patients believed their 
health insurance covered reasonable health care costs for medically beneficial 
care.4 At least part of the support for the latter belief lies in the typical langu­
age of health insurance contracts specifying that the scope of coverage encom­
passes reasonable and necessary care. Specifically limited health insurance, 
such as dental insurance, quite clearly would not create the legitimate expecta­
tion of more general coverage. The difficult question is whether any off-label 
therapy falls within the general scope language extending coverage to care that 
is reasonable and necessary. The first set of beliefs-beliefs about the likely 
benefit to be obtained from off-label therapy-is central to answering this 
question, because the argument that the care is reasonable and necessary rests 
on the likely benefit to be obtained from it. Thus patients' beliefs-that care 
would be beneficial and thus covered-were crucial to the purpose of the 
statutes. But are these beliefs reasonable? If not, what conclusions should we 
draw about the statutory mandates? Do the mandates merely bootstrap unrea­
sonable original expectations into a new set of expectations based in the 
statutes themselves? Or, do the statutes reinforce expectations that were 
originally legitimate? 

The reasonableness of beliefs is a complex matter. Objective reason­
ableness is linked to whether the beliefs are true or false. Respect for persons 
as choosers and planners of their lives, however, must be based on the world 
as they see it, not as it objectively might appear from a detached perspective. 5 

Beliefs that are founded upon the best available evidence may eventually be 
discovered to be false. In addition, people may hold beliefs that happen to be 
true but that are accepted for remarkably irrational reasons. 

Respecting autonomy requires respecting choices based upon beliefs 
that seemed well founded when the choice was made, even if the beliefs might 
later be rejected. Conversely, respecting autonomy does not require honoring 
irrationally accepted expectations or expectations ba8ed on beliefs that are 
known to be false or that reasonably should be regarded as suspect. Auton­
omy is reasoned choice, not unfounded choice. To the same effect is the con­
sequentialist concern that honoring expectations based upon beliefs reasonable 
people should know are false may create counter-productive incentives for un­
reasonable reliance. Subjective reasonableness is thus more relevant to 
expectations than objective reasonableness. But what does subjective reason­
ableness demand in areas of life where people are likely to be unreasonable-

3. In the language of these legislatures. "Denial ofpaymentforoff-label use can interrupt 
or effectively deny access to necessary and appropriate treatment for a person being treated for 
a life-threatening illness." ALA. CODE§ 27-1-10.1(3) (2003); N.J. STAT. ANN.§ 26:1A-36.9(d) 
(West 2003); TENN. CODE ANN.§ 56-7-2352(4) (2003). 

4. The statute might simply have mandated the coverage. To say that the coverage is 
"legitimately expected" suggests that patients believe it will be forthcoming, and that these 
beliefs ought to be honored. Using the language of expectations suggests that what people 
believe plays a role in justifying the coverage. 

5. See THOMAS NAGEL, THE VIEW FROM NOWHERE (1986). 



218 INDIANA HEALTH LAW REviEW [Vol. 1:213 

areas.that are momentous, but difficult to understand and emotionally chal­
lenging? 

Health care is clearly such an area. Even leaving emotions aside, health 
care decisions often require significant scientific and technical expertise. 
Medical care, especially complex care for diseases that are serious and diffi­
cult to treat, is a context in which expert advice appropriately looms ·large. 
Patients' beliefs about the likely efficacy of care are more reasonable when 
they are based on the recommendations of recognized experts-for example, 
physicians in centers designated as comprehensive cancer centers by the 
National Cancer Institute (''NCf').6 

In addition, health care is not an area where emotions can be left aside; 
patients with life-limiting illnesses are understandably frightened and vulner­
able. Moreover, such patients may be in a state of mind in which their powers 
of independent understanding are not as fully operative as they might be in 
other circumstances, such as understanding architectural advice in home­
building. In facing serious illness, patients struggle to understand difficult 
scientific concepts and understandably rely on expert advice. A patient's 
belief that therapy is likely to be safe and effective, when reached after 
consultation with a reasonable expert, is thus subjectively reasonable. This 
is not to say that all expert advice is, in turn, reasonable. How, then, does a 
judgment about the reasonableness of experts' beliefs affect a judgment about 
the reasonableness of a patient's beliefs? 

One important factor regarding a patient's belief is the degree to which 
the experts' beliefs have scientific support. Available evidence may range 
from promising results in a few treated patients-results that may be sugges­
tive but may also be limited to certain patients-to a weU-designed, random­
ized clinical trial.7 Moreover, experts may be swayed by enthusiasm gener­
ated by genuinely promising results in the most recent patients treated. 
Experts may also be affected by financial interests in the research they are 
conducting. 8 At a minimum, experts should be aware of the level of evidenti­
ary support for their claims and any potential for conflict of interest. They 
should also attempt to make these factors transparent to patients. 9 

6. See Nat'l Cancer Inst., The National Cancer Institute Cancer Centers Program, at 
http://cis.nci.nih.gov/fact/t .. 2.:htm Oast visited July 25, 2003) (describing this designation, 
including the requirement for rigorous peer review). 

7. See Executive Comm. Working Group, Medicare Coverage Advisory Comm., 
Recommendations for Evaluating Effectiveness (Feb. 23, 2001), available at http://www.cms 
.hhs.gov/mcac/8b1-i9.asp (discussing different levels of evidentiary support) [hereinafter 
Medicare Coverage Advisory Comm.]. 

8. See Karine Morin et al., Managing Conflicts of Interest in the Conduct of Clinical 
Trials, 287 JAMA 78 (2002) (discussing how conflicts of interest may affect the conduct of 
clinical trials). 

9. The transparency model of informed consent has been explored and defended by 
Howard Brody. The basic idea is that the physician's reasoning should be made clear 
(transparent) to the patient as part of the process of informed consent. See Howard Brody, 
Transparency: Informed Consent in Primary Care, 19 HAsTINGs CENTER REP. 5 (1989). 



2004] MANDATED COVERAGE OF EXPERIMENTAL THERAPY 219 

Experts, however, may fail to explain these differences carefully. 
Physicians have difficulty admitting and explaining uncertainty to patients.10 

Recent studies of patients' understanding of medical research indicate that 
roughly half do not even understand that they are participating in research. 11 

When a physician describes therapy in terms of chances of cure or partial 
remission, the patient may easily misunderstand. For example, informed con­
sent forms for participation in clinical trials often begin with a statement that 
the trial is for the purpose of research but then continue to include treatment 
language: ''you will begin treatment on day one" and ''treatment will continue 
until your disease progresses," rather than ''your entry into the research will 
begin on day one," and ''your participation in the study will continue until 
your disease progresses." Consent forms may include a schedule of evalua­
tion extending over many years or ''for life," when in fact the expected end­
point of the research is only a. few months of event-free survival. 

Another factor confounding judgment of subjective reasonableness is 
how people frame information about risk. Tversky, Kahneman, and other · 
students of decision theory have demonstrated that the description of a 
decision problem affects choice. Faced with the same gambles described 
differently, people do not consistently elect the. option with the maximum 
expected utility.12 That is, they do not behave rationally, if rationality is 
understood in terms of utility maximization. Instead, they are risk averse 
when offered the option of an assured gain versus a lottery with a chance of 
no gain but a higher expected payoff vaiue. At the same time, people are risk 
acceptant when offered the parallel option of an assured loss versus a lottery 
with a chance of no loss but a higher expected negative payo:ff.13 Psycho­
logically, people want to gain something but lose nothing; whether a decision 
is framed as a potential loss or a potential gain makes much of the difference 
in how they choose.14 Tversky and K.ahneman further argue that these and 
other descriptive findings about how people make choices should be of 
interest to the normative theory of rational choice.15 Tversky and others have 
demonstrated this phenomenon in health care in the context of the choice 
between surgery and radiation for lung cancer.16 Arguably, then, experts 

10. See JAY KA1Z, THE SILENT WoRlD OF DocToR AND PA'rmNT (1984). 
11. Nancy E. Kass et al., Trust: 11ae Fragile Foundation of Contemporary Biomedical 

Research, in INTERVENTION ANDR.ERBcriON 494-99 (Ronald Munson ed., Wadsworth 6th ed. 
2000). 

12. Daniel Kahneman & Amos Tversky, Choices, Values, and Frames, in CHoiCES, 
VALUES, AND FRAMEs 1-16 (Daniel Kahneman & Amos Tveisky eds., 2000). 

13. Id. 
14./d. 
15. /d. at 209-23. 
16. B.J. McNeil, S.C. Pauker, H.C. Sox, Jr., & A Tversky, On the Elicitation of 

Preferences for Alternative Therapies, 306 NEW ENG. J.MED. 1259 (1982). Surgery involves 
a chance of death during the actual therapy, unlike the alternative of radiation. /d. However, 
the overall survival data are better for surgery. Id. Patients were less likely to opt for surgery 



220 INDIANA HEALTH LAw REviEw [Vol. 1:213 

advising patients should be aware of these framing effects and should be 
discouraged from using them as a strategy to influence patients' choices.17 In 
sum, reliance on experts, by itself, does not undermine the reasonableness of 
patients' beliefs. However, for the reliance to be reasonable, experts should 
strive to be clear about evidentiary support and about their own possible 
conflicts of interest, and also should strive to not play upon known factors 
affecting patient misperceptions. 

Although reinforcement likely starts with reliance on expert advice, as 
described above, general social reinforcement arguably also affects beliefs 
about medical care. For example, newspaper reports of promising new 
therapies, rumors that a "cure" is just around the comer, or anecdotal reports 
of success from neighbors or friends may all feed unrealistic expectations 
about what is likely or possible from medical research. It is difficult to disen­
tangle the respective influence of such reinforcement, much less determine 
whether reliance on it is reasonable. 

The simplest claim about reasonableness would be that the more 
authoritative the source of reinforcement, the more reasonable the belief. 
Thus, scientific reporting in newspapers such as the New York Times, 
information on reputable web sites such as that of the NCI, official reports 
from government agencies, and reinforcement found in statutes, administrative 
regulations, and judicial decisions would be considered more authoritative 
than, for example, a neighbor's account of a supposedly promising new 
therapy. For this reason, the role of legal mandates in creating expectations 
about entitlement to medical care warrants particularly careful scrutiny. Part 
five (V) below, however, argues that legislatures and courts have not been as 
careful as they should be about the expectations they have created. 

All of these factors affecting subjective reasonableness-the vulner­
ability of patients, the importance of framing effects, the role of experts, and 
social reinforcement-converge in the situation of a patient with life-limiting 
cancer. Social policy and law pose additional complications in the case of 
cancer treatment. Patient advocacy groups and feminist groups have pressed 
the sympathetic situations of patients with breast cancer through language that 
assumes therapeutic efficacy for interventions such as high-dose chemo­
therapy with stem cell support. In response, a number of states have adopted 
coverage mandates for patients in clinical trials of new modalities of cancer 

if the choice of therapy was posed in terms of risks of mortality rather than in terms of chances 
of survival. /d. 

17. Other characteristic features of decision-making are also identified by work in the 
Tversky and Kahneman tradition. For example, the understanding of probability is not linear; 
people are far more influenced by the differences between 0 and 1% and between 99 and 100%, 
than they are by other differences of 1% on the probability scale. This feature may be especially 
important when people additionally confuse a certainty ("curej with a probability ("chance of 
remission") in medical decisions. See, e.g., Daniel Kahneman & Amos Tversky, Prospect 
Theory: An Analysis of Decision Under Risk, in CHOICES, VALUES, AND FRAMEs, supra note 
12, at 17-43. 



2004] MANDATED COVERAGE OF ExPERIMENTAL THERAPY 221 

treatments (see Part IV). But these coverage mandates may outstrip the 
justifications for them based solely on therapeutic efficacy (see Part m. Even 
more in the public eye, breast cancer patients (or their estates) have won high­
profile lawsuits challenging their insurers for denying coverage for high-dose 
chemotherapy with stem cell support (see Part V). These cases foster the 
assumption that the care would be life-saving, and the concomitant conclusion 
that such care should be provided as a matter of entitlement. 

II. HIGH-DOSE CHEMOTHERAPY WITH STEM CELL SUPPORT FOR 
BREAST CANCER: THE CLINICAL STORY 

A. The Therapeutic Strategy 

High-dose chemotherapy has been explored as a therapeutic option for 
a number of different types of cancer. This type of therapy has the advantage 
of potentially eradicating cancers that cannot be treated successfully with 
lower doses of chemotherapy.18 The disadvantages include cost-which may, 
depending on the patient's condition and complications of therapy, approxi­
mate $250,00019-and toxicity. High-dose chemotherapy, administered 
systemically, will kill all cells in the body that divide more rapidly than others, 
including bone marrow cells (the clinical term for destruction of bone marrow 
cells is "myoablative"). Without functioning bone marrow, patients are 
unable to produce new blood cells and will die from infection, hemorrhage, 
or anemia. Accordingly, patients receiving high-dose chemotherapy must be 
supported by are-infusion of cells that can recreate functioning bone marrow. 
During the period before bone marrow function returns, such patients are at 
risk of infection and other complications; the risk of death in the period of 
administration of high-dose chemotherapy and immediately thereafter is as 
high as ten percent, depending on the patient population. 20 

If the patient donates the re-infused cells, the transplant is autologous. 
If a matched donor provides the cells, the transplant is allogeneic. The re-in­
fused cells may be harvested bone marrow cells or circulating stem cells. 
Autologous transplants avoid graft-versus-host disease and have a lower rate 
of death associated with the immediate transplant procedure; however, 
autologous transplants may be less successful long term because of the 

18. For a good description of high dose chemotherapy with stem cell support, see Nat' I 
Cancer Inst., Bone Marrow Transplation and Peripheral Blood Stem Cell Transplantation: 
Questions and Answers, at http://cis.nci.nih.gov/fact/7 _ 41.htm (last visited Feb. 5, 2004 ). 

19. For a good discussion of the direct and indirect costs of bone marrow transplantation, 
see NAT'LBONE MARROW TRANSPLANT LINK, A REsOURCE GuiDE FOR BONE MARROW/STEM 
CELLTRANSPLANT(200l),athttp://www.nbmtlink.org/rg.asp#cost(lastvisitedMar.29,2004). 

20. INT'LMYELOMAFOUND., UNDERSTANDING AUTOLOGOUS BONE MARROW AND STEM 
CELL TRANSPLANTATION: HERE'S WHAT You SHOUlD KNOW 23, at http://myeloma.org/ 
images/content_pdf/BMT.pdf (last visited Mar. 29, 2004). 



222 INDIANA HEALTH LAW REviEW [Vol. 1:213 

possibility the disease will not be fully eradicated.21 Very recently, some 
transplant programs have been exploring lower-dose chemotherapeutic 
regimes coupled with allogeneic transplants. These programs are called mini­
transplants. Higher-risk patients may be better able to tolerate the chemo­
therapeutic regimens involved. Additionally, there is some hope that the 
mixed chimerism of the patient's own bone marrow, not fully ablated by the 
chemotherapy, together with the donor marrow, will create an immune 
reaction of independent therapeutic efficacy.22 

In the 1980s, a number of transplant programs began exploring high­
dose chemotherapy with transplant support for the treatment of metastatic 
breast cancer. Early reports suggested the therapy might be effective, al­
though these reports came from smaller (Phase m trials that did not randomize 
patients and that did not report survival data over the long term. 23 The pro­
cedure came to be seen as life-saving, even miraculous. One court's opinion 
in a challenge by a patient to her insurer's denial of coverage was typical: 

In light of the medical testimony that HDC-ABMT [high­
dose chemotherapy-autologous bone marrow transplant] is an 
effective form of therapy for breast cancer and that breast 
cancer patients who earnestly seek the treatment experience 
better prognoses than those who choose conventional treat­
ments, we find that the trial court did not err in fmding that 
HDC-ABMT is not experimental or research in nature .... 24 

Other courts credited the therapy with lengthening life or improving the 
quality of life, despite very short durations of survival for patients.25 This 
enthusiasm, however, has been dashed by subsequent research. Recent studies 
indicate no long-term or relapse-free survival benefit associated with high­
dose chemotherapy and stem cell support for breast cancer patients generally, 
although research is ongoing. 26 The NCI web site designed for health care 
providers opines that "[i]n the absence of data suggesting.a benefit from high-

21. See generally id. (discussing autologous bone marrow transplants for multiple 
myeloma). 

22. See, e.g., Duke Univ. Health Sys., About the Procedure: Bone Ma"ow Transplant 
& Stem Cell Transplant Program, at http://bmt.mc.duke.edulprocedure/t-minitransplant.asp 
(last visited Mar. 29, 2004). 

23. See, e.g., W.P. Peters et al., High-Dose Combination Alkylating Agents with Bone 
Marrow Support as Initial Treatment for Metastatic Breast Cancer, 6 J. CUNICALONCOLOGY 

1368 (1988) (study at Duke University). 
24. Taylor v. Blue Cross/Blue Shield of Mich., 517 N.W;2d 864, 869-70 (Mich. Ct. App. 

1994). 
25. See, e.g., Scalamandre v. Oxford Health Plans, Inc., 823 F. Supp. 1050 (E.D.N.Y 

1993) (The patient had the procedure performed in September and October and then died the 
following January). 

26. For a good overview, see Astrid Mayer & Helena Earl, Whither High-Dose Chemo­
therapy in Breast Cancer? 3 BREAST CANCER REsEARCH 8 (2001). 



2004] MANDATED COVERAGE OF EXPERIMENTAL THERAPY 223 

dose chemotherapy with stem cell support, this remains an area of clinical 
evaluation.'m Toward this end. several clinical trials are ongoing. 28 The trials 
vary by chemotherapeutic regimen; some trials are of higher dose regimens 
and others of mini-transplants. These clinical trials have stringent criteria for 
entry, including patient age, disease stage, and prior treatment history; they are 
designed more for research than treatment purposes. 

Nationwide, sophisticated patients seeking information about cancer 
treatment have increasingly turned to the Internet Patients who explore web 
sites of major cancer centers, however, will encounter significantly different 
views of the current status of myoablative therapy with stem cell support for 
breast cancer. An August 2003 survey of the web sites of the forty-two 
centers designated as Comprehensive Cancer Centers by the National 
Institutes of Health ("Nlll'') reveals that twenty-four of these web sites feature 
readily accessible, extensive, and clear descriptions of the difference between 
a clinical trial and treatment.29 On some web sites, however, the description 
of the difference is difficult to find30 or requires following a link to another 
site.31 Some of these web sites even list high-dose chemotherapy with stem 
cell support as a possible treatment option for breast cancer. The Lombardi 
Cancer Center at Georgetown University features a "success story,. of a 
patient who survived breast cancer through a bone marrow transplant 32 The 
City of Hope in Los Angeles boasts 5200 bone marrow transplants performed, 
the third largest number in the United States. 33 Its site includes testimonials 
of two patients with advanced breast cancer. The web site description does 

27. Nat'l Cancer Inst., Stage 111 B, 111 C, N. Recurrent, and Metastatic Breast Cancers: 
Cytotoxic Chemotherapy, available at http:/lwww.cancer.gov/cancerinfo/pdq/treatmentlbreast/ 
healthprofessionali#Section_278 (last visited Nov. 24, 2003) (citing DA Berry et al., High­
Dose Versus Standard Chemotherapy in Metastatic Breast Cancer: Comparison of Cancer and 
Leukemia Group B Trials with Data from the Autologous Blood and Marrow Transplant 
Registry, 20 J. CUNICAL0NCOLOGY 743 (2002)). 

28. Nat'l Cancer Inst., Clinical Trials, available at www.cancer.gov/clinicaltrials/ (last 
visited Mar. 16, 2004). The site is updated frequently to indicate the number of open trials of 
high-dose chemotherapy in which breast cancer patients are included. The listed trials vary by 
patient population, type of chemotherapy or radiation therapy, and therapeutic intensity. 

29. Data on file with the author and the Indiana Health Law Review. 
30. See, e.g., The Abramson Cancer Ctr. of Univ. of Pa., at www.oncolink.com (last 

visited Nov. 24, 2003). 
31. See, e.g., Univ. of N.C. Lineberger Comprehensive CancerCtr., at http://cancer.med 

.unc.edu/patient/default.asp (last visited Jan. 1, 2004). 
32. Lombardi Cancer Ctr., Success Stories: Surviving Cancer, at http://lombardi .george 

town.edu/success/surviving.htm (last visited Jan. 1, 2004). 
33. City of Hope, at http://www.cityofhope.org (last visited Nov. 24, 2003) (providing 

many success stories of treating breast cancer patients with bone marrow transplantation). Other 
comprehensive centers with web sites which suggest that bone marrow transplantation might 
be available as a treatment for breast cancer include the Norris Center at University of Southern 
California, at http://ccnt.hsc.usc.edu/serviceslbreast_cancerltreatment.htm (last visited Nov. 24, 
2003), and the Chao Center at University of California-Irvine, at http://www.ucihealth.com/ 
cancer/index.htm (last visited Nov. 24. 2003). 



224 INDIANA HEALTII LAw REVIEW [Vol. 1:213 

note that seventy-five percent of the cases treated at the City of Hope involve 
leukemia or lymphoma (cancers in which the bone marrow itself is a cancer 
site), but lists breast cancer as an "other malignancy" treated. 34 The Roswell 
Park Cancer Institute at State University ofNew York-Buffalo lists autologous 
bone marrow transplantation as a treatment option for "select" patients. 35 The 
M.D. Anderson Cancer Center web site contains a careful description of 
clinical trials, their risks, and the possibility that insurance will not· pay for 
them.36 A separate location on the web site lists all of the clinical trials on­
going at the center, without the implication that clinical trial participation is 
indicated therapy. It also, however, lists myoablative therapy with stem cell 
support as a treatment option for advanced breast cancer. The Fred Hutchin­
son Cancer Research Center in Seattle contains a glowing description of a 
reunion of bone marrow transplant patients, which includes this text: 

Malignant diseases such as breast cancer and others are now 
being treated in this way, since the same general principles of 
destroying malignant cells with mega doses of drugs and 
radiation are used. Responses in breast cancer vary, but more 
time is necessary to evaluate the long-term results. Some 
autoimmune diseases are also now the subjects of treatment 
with stem-cell transplants. 37 

The Mayo Clinic, by contrast, does not include breast cancer as a solid tumor 
treated by bone marrow transplantation. 38 

A fair inference from this survey is that patients accessing the web sites 
of these designated comprehensive cancer centers may be at best confused 
about whether myoablative chemotherapy with stem cell support is currently 
regarded as standard, effective treatment for advanced breast cancer. Under­
standably, patients wanting the best medical care may conclude that because 
some centers offer this therapy as effective treatment, it is an option they 
should pursue and an option others should support as life-saving treatment for 
them. 

34./d. 
35. Roswell Park Cancer Inst., Breast Center, at http://www.roswellpark.org!Document 

.asp?lid=723 (last visited Nov. 24, 2003). 
36. The Univ. of Tex. M.D. Anderson CancerCtr., athttp://www.mdanderson.org/ (last 

visited Nov. 24, 2003). A number of other cancer centers also have designed their web sites to 
list clinical trials separately. 

37. E.D. Thomas & D. Thomas, Fred Hutchinson Cancer Research Ctr., The Promising 
Future of Transplantation: Expanded Application of Hematopoietic Transplantation, at 
http://www.thcrc.orglvisitor/patient_experience/reunion.html (last visited Nov. 24, 2003). 

38. Mayo Clinic, at http://www.mayoclinic.org/bmarrowtransplant-rstl (last visited July 
25,2003). 



2004] MANDATED COVERAGE OF EXPERIMENTAL THERAPY 225 

B. Insurance Coverage for Myoablative Chemotherapy for Breast Cancer 

In the private market, health insurance coverage is established by con­
tract. Insurance contracts typically provide coverage for medically indicated 
therapy, with enumerated exclusions and patient payment responsibilities. 
Contract language will likely exclude therapy that is "experimental" or 
"investigational" in the judgment of the insurer. Patients faced with coverage 
denials may challenge these judgments in court. Because more than eighty­
five million patients are insured by its plans, Blue Cross and Blue Shield 
("BC/BS") Association has been a leader in assessing the evidence supporting 
medical practices.39 The BC/BS Technology Evaluation Center (''TEC"), 
established in 1985 and working in collaboration with Kaiser Permanente 
since 1993, provides evidence-based evaluations.40 TEC does not list an 
evaluation of myoablative therapy with stem cell support for breast cancer, 
although it does for other cancers, such as multiple myeloma. In the case law, 
courts have expressed suspicion about technology assessment organizations 
and instead have relied on the testimony of treating oncologists. 41 

Medicare beneficiaries include both those qualifying for Social Security 
because they have worked the requisite forty quarter-years and those qualify­
ing because they are disabled. Like BC/BS, Medicare often plays a major role 
in coverage determinations because it covers over forty million patients and 
invests a high level of technical scrutiny in coverage decisions.42 Under the 
Medicare statute, one criterion for coverage determinations is whether the 
intervention is ''reasonable and necessary.'t43 As of this writing, cost is not 
listed as a factor in Medicare regulations governing judgments of reason­
ableness and necessity,44 although discussion of this issue may resume with 
the public policy spotlight on the costs of the Medicare program. Coverage 
decisions are based on whether data supports the efficacy of therapy in the 
population over the age of sixty-five or in other patients eligible for 
Medicare.45 A difficulty facing some coverage decisions is that there is little 
data about the safety or efficacy of some therapies in patients over the age of 
sixty-five. Myoablative chemotherapy is an example because of its toxicity. 
Some of the clinical trials now underway involving myoablative therapy for 

39. Blue Cross Blue Shield Ass'n, at http://www.bcbs.com/tecfmdex.html (last visited 
July 25, 2003). 

40. Id. 
41. See,e.g.,Adamsv.BlueCross/BlueShieldofMd., 757F.Supp. 661 (D.Md.1991); 

Pirozzi v. Blue Cross-Blue Shield of Va., 741 F. Supp. 586 (E.D. Va. 1990). 
42. Ctts. for Medicare & Medicaid Servs., at http://www.cms.gov/medicare/ (last visited 

Mar. 9, 2004) [hereinafter CMS web site]. 
43. 42 U.S.C. § 1395y(a)(l)(A) (2003). 
44. For the regulatory definition of the scope ofMedicare covered services, see 42 C.F.R. 

§ 402.3 (2003). 
45. Medicare Coverage Advisory Comm., supra note 7. 



226 INDIANA HEALTH LAW REVIEW [Vol. 1:213 

breast cancer specify an upper age limit of sixty and thus will provide data that 
is at best indirectly relevant to the majority of the Medicare population.46 

Medicare coverage decisions may be made nationally or by local carriers 
if there is not a national coverage determination. 47 There is a national cover­
age determination that autologous stem cell transplantation is not covered for 
solid tumors, including breast cancer. The coverage decision was that there 
is insufficient evidence to show that high dose chemotherapy plus stem cell 
support is reasonable and necessary in the Medicare population. 48 No national 
coverage determination has been issued for allogeneic transplants, but these 
transplants are not listed as covered services in the Medicare population for 
patients with solid tumors.49 Nor has there been a national coverage deter­
mination for mini-transplants. 5° 

ill. INSURANCE COVERAGE FOR PATIENTS PARTICIPATING IN 

CUNICAL TRIALS 

Clinical trials are designed to evaluate whether new medical interven­
tions are superior to standard patient management, including both treatment 
and diagnosis. The current canonical view is that clinical trials are not ethical 
without "equipoise:" there must be no good reason to believe that the innova­
tion to be tested is either better or worse than standard management of the 
patient. 51 Without equipoise, the trial would be unethical because the patients 

46. Examples are the trial of Allogeneic Peripheral Stem Cell Transplantation After Anti­
thymocyte Globulin, High-Dose Melphalan, and Fludarabine in Treating Women with Meta­
static Adenocarcinoma of the Breast, at http://clinicaltrials.8ov/ct/show/NCT0007 4269?order 
=4 (last visited Mar. 8, 2004), and the trial of Chemotherapy and Stem Cell Transplantation in 
Treating Patients with Stage /liB Breast Cancer, at http://clinicaltrials.gov/ct/show/NCTOOOO 
3042?order=7 (last visited Mar. 8, 2004). 

47. Information about how Medicare coverage determinations are made is available on 
the CMS web site, at http://www.cms.gov/coverage/8a.asp (last visited Mar. 8. 2004) (especially 
the June 4, 2002 Report to Congress available on the web site). 

48. The full list of national coverage determinations can be found on the CMS web site, 
at http://www.cms.gov/mcdlindex_Iist.asp?Iist_type=ncd (last visited Mar. 8, 2004). 

49. This information is from the CMS web site, at http://www.cms.hhs.gov/ncdlsearch 
display.asp?NCD_ID=45&NCD_ vrsn_num=4 (last visited July 25, 2003). A number oflocal 
carriers have made parallel determinations that stem cell transplants are not covered for solid 
tumors in patients in the Medicare population. !d. These include the Wisconsin Physicians 
Service Insurance Corporation (for Wisconsin, illinois, Michigan, and Minnesota), Palmetto 
Government Benefits (Ohio and West Virginia), Blue Cross and Blue Shield of Nebraska 
(Nebraska; also clarifies that "rninitransplants" are not covered for solid tumors), Group Health, 
N.Y. (New York, Queens), HGS Administrators (Pennsylvania), First Coast Service Options 
(Florida), and Blue Cross and Blue Shield of Arkansas (Arkansas, New Mexico, Oklahoma, 
northeastern and southern Missouri, and Louisiana). !d. 

50. CMS web site, at http://www.cms.gov/mcdlviewncd.asp?ncd_id=35-30.1&ncd_ 
version=5&show=all (last visited Mar. 8, 2004). 

51. Benjamin Freedman, Equipoise and the Ethics of Clinical Research, 317 NEW ENG. 
J.MED. 141 (1987). 



2004] MANDATED COVERAGE OF EXPERIMENTAL THERAPY 227 

in one arm of the trial would be getting care already reasonably believed to be 
less good than care in one of the other arms of the trial. "Equipoise," 
however, may be at best an unattainable ideal and at worst the basis of mis­
conceptions. Although it is not ethical to design a clinical trial that withholds 
management of known value, many clinical trials approved today appear to 
comport uneasily with the idea of equipoise. Fairly strong suspicions may be 
present among researchers that one arm of the trial is preferable to others, but 
unless the trial is stopped because of an adverse event or a significant result, 
it will continue. Many trials include a placebo arm, albeit infrequently in 
cancer trials. Even when there is a generally accepted standard therapy, 
placebo arms may be included if the risks to the patient are minimized and 
there is reason to believe there may be a large placebo effect or that the 
standard therapy carries disadvantages of its own. Many trials also include a 
washout period in which a patient discontinues existing medications, again if 
the risks to patients are minimized. These concerns about equipoise have led 
at least two authors to argue that the ethics of research and the ethics of 
therapy should be detached, lest research be confused with the most effective, 
patient-centered therapy.52 

Randomized clinical trials are considered to give the best information 
about the efficacy of a given medical intervention. 53 In such clinical trials, 
patients are selected according to entry criteria designed to elicit reliable and 
valid scientific information about the intervention being studied. 54 Too much 
variance in patient characteristics, for example, might obscure whether 
differences in patients or in clinical management explain differences in out­
comes. In the best clinical trials, patients are randomized among study arms 
to ensure that the variation being studied, and not variations among patients, 
is the relevant independent variable. 55 

Clinical trials of new therapeutic regimens are classified in "phases. "56 

The primary goal of a Phase I trial is evaluation of safety, safe dosage ranges, 
and side effects. Therapeutic efficacy is only a secondary consideration. 
Phase I trials of cancer chemotherapy regimens generally enroll a small 
number of patients, from twenty to eighty. Patients are evaluated first at 
lower, possibly sub-therapeutic, dosages of the proposed regime. If patients 
at the lower dosage do not experience significant toxicities, patients are 
enrolled at higher dosage levels. Dose escalation ceases when two patients 
experience significant toxicities. Patients enrolled in Phase I trials generally 

52. Franklin G. Miller & Howard Brody, A Critique of Clinical Equipoise: Therapeutic 
Misconception in the Ethics of Clinical Trials, 33 HASTINGS CENTER REP. 19 (2003), available 
at2003 WL 17821782. 

53. ld. 
54./d. 
55. Medicare Coverage Advisory Comm., supra note 7. 
56. See generally Nat'l Inst. Health, CancerTrials.gov, at www.clinicaltrials.gov (last 

visited Jan. 1, 2004) (providing a general overview of clinical trials and discussing phases). 



228 INDIANA HEALTH LAW REVIEW [Vol. 1:213 

have exhausted other available therapies; therefore, it is considered ethical to 
offer them the possibility of benefit even if the trial's main goal is to establish 
tolerated dosage levels. Phase ll trials randomize patients between the investi­
gational regimen and a standard therapeutic regimen. Enrolling larger num­
bers of patients, from 100 to 300, the trials are designed to see whether there 
are statistically significant differences between treatment arms and whether 
levels of toxicity are acceptable. Phase ll trials focus on determining short­
term efficacy and safety of therapy. These trials have stopping rules under 
which they will be ended if reported toxicities are unacceptable or significant 
differences in efficacy become apparent. Phase ill trials enroll a larger 
number of patients, typically from 1000 to 3000, and are undertaken only if 
Phase ll trials suggest further study is warranted. Phase m cancer trials 
typically require interim data analyses. They are stopped if there is clear 
evidence that one treatment arm is significantly preferable to others and, 
therefore, there is no longer clinical equipoise between the arms of the trial. 

Phase N trials undertake continued collection of data after a new drug 
is given marketing approval based on data from earlier trials. The goal of 
Phase N is to collect data on an ongoing basis as an approved therapy be­
comes employed in the general population of patients in need of treatment. 
Distribution of a therapy into the general population of patients, outside the 
research context, may reveal quite different aspects of the therapy's risks and 
benefits. 

Thus, the balance between scientific and therapeutic goals shifts with 
the phase of a clinical trial. For Phase I trials, therapeutic goals are clearly 
secondary, although the investigational regimen would not be under study 
without some reason to believe it holds therapeutic promise. For Phase N 
trials, therapeutic goals are undoubtedly primary, with the secondary aim 
being the ongoing monitoring of safety and efficacy. However, it is not at all 
clear that patients understand the differences between these types of clinical 
trials, much less the difference between research and therapy itself. 

The potential for confusion between therapy and research is so signifi­
cant that the confusion has been given a label among writers in bioethics: the 
"therapeutic misconception."57 If patients believe they are receiving what 
their physician regards as the "best treatment" for them, yet they are partici­
pating in a clinical trial in which they are randomized among treatment arms 
believed to be in equipoise, they fall under the therapeutic misconception. 58 

The confusion between research and therapy is of course most significant in 
Phase I trials, where the dosage level the patient receives may be set to 
evaluate something other than therapeutic efficacy. Patients assigned to the 
initial dosage levels, for example, may receive less efficacious therapy, but 

57. Miller & Brody, supra note 52. 
58. ld. 



2004] MANDATED COVERAGE OF EXPERIMENTAL THERAPY 229 

may also avoid more significant toxicities. 59 The misconception is also pre­
sent in Phase III trials, where patients may be randomized among treatment 
arms. In theory, Phase III trials are designed to be in equipoise; it would be 
unethical to expose a patient to the possibility of assignment to a treatment 
arm known to be less efficacious than other treatments. On the other hand, 
Phase III trials are undertaken against the background of suspicion that one 
arm will prove to be superior to the other(s). This background may be either 
inconclusive evidence that a new regimen is superior (e.g., evidence from a 
smaller Phase ll trial), evidence that there are efficacy or toxicity problems 
with standard therapy, or other advantages to the therapy under investigation. 
Phase III trials are conducted in delicate balance between the need for data 
that are sufficiently significant to demonstrate the superiority of one treatment 
arm and the maintenance of sufficient clinical equipoise to permit continua­
tion of the study. 

The therapeutic misconception is fostered by the fact that patients with 
cancer face life-limiting illness. Patients are in situations of significant 
personal, physical, and frequently even financial vulnerability. By the time 
patients contemplate entering a clinical trial, they may have already attempted 
and, in the language of medicine, "failed" other available treatment options. 
Entry into a clinical trial may be an option of last resort. 60 One important 
study suggests that patients entering clinical trials do not adequately under­
stand that they are engaged in research or what being engaged in research 
actually means. 61 Although ninety percent of patients in the study expressed 
overall satisfaction with the informed consent process, the majority (74%) did 
not understand that they were receiving non-standard treatment.62 Many also 
did not understand the potential for additional risk presented by participation 
(63%), did not understand that the treatment was unproven (70%), did not 
understand the uncertainty of benefits to self (29% ), and did not realize that 
trials are done mainly to benefit future patients (25% ).63 In this study, 
increased knowledge was associated with college education, English as the 
only spoken language at home, the use of the United States NCI consent form 
template, signing the consent form later than at initial discussion, presence of 
a nurse, and careful reading of the consent form. 64 Other researchers have 

59. /d. 
60. See Benjamin Freedman, A Response to a Purported Ethical Difficulty with 

Randomized Clinical Trials Involving Cancer Patients, 3 J. CUNICAL ETHICS 231 (1992) 
(discussing issues regarding patients with cancer in clinical trials); Maria Merritt, Moral 
Conflict in Clinical Trials, Address Before the Pacific Division Meeting for the American 
Philosophical Association, Seattle, Wash. (Mar. 2002) (on file with the author) (discussing 
tensions between clinical duties and scientific duties in Phase I trials). 

61. Steven Joffe et al., Quality of Informed Consent in Cancer Clinical Trials: A Cross-
Sectional Survey, 358 LANCET 9295 (2001). 

62. ld. 
63. ld. 
64. ld. 



230 INDIANA HEALTH LAw REVIEW [Vol. 1:213 

documented that patients from racial and ethnic minorities have particular 
difficulties understanding clinical trials.65 

Nonetheless, patients in many clinical trials, even Phase I trials, pro­
bably do receive benefits. Patients are monitored closely, perhaps more 
closely than they would be outside of the trial setting. Clinical trials spon­
sored by the NIH or by major cancer networks are designed on the basis of 
state-of-the-art knowledge about available options for patients. The treatment 
teams involved in such trials are likely to be highly experienced and may be 
better able to manage side effects and disease symptorns. Finally. drugs or 
medical devices that have not yet received market clearance from the Food 
and Drug Administration ("FDA") will typi~ally be provided free of charge, 
or in some cases at the cost of production, which may be a significant cost 
savings.66 The advantages are well documented for cancer therapy involving 
children, where participation rates are far higher than the estimated three to 
five percent rate of participation for adults with cancer.67 The results demon-

. strate higher success rates in treating children with cancer and better develop­
ment of cancer therapies for children· compared to cancer treatment for 
adults.68 

Despite the fact that clinical trials involving patients with cancer 
typically offer experimental drugs without charge, patients (or their health 
insurers) generally are expected to pay all other charges of treatment received 
in the clinical trial. 69 Patients are billed for the costs of administering the 
drug; the costs of diagnostic studies required to stage disease or assess disease 
progression; the costs of any hospitalizations, radiation therapy, or surgery; 
and the costs to treat any side effects or adverse reactions to therapy.70 The 
exceptions to these patient responsibilities are the costs of data collection or 
assessments required solely for the study itself and for monitoring therapy.71 

Patients entering clinical trials are advised that they will be responsible for 
costs, often without accurate estimates of what those costs might be because 
of expenses related to side effects, adverse reactions, or patients' responses to 

65. Elmer E. Huerta & Everly Macario, Communicating Health R~k to Ethnic Groups: 
Reaching Hispanics as a Case Study, 25 J. NAT'L CANCERINST. MONOGRAPHS 23 (1999). See 
also L. Ellington et al., Decision Making Needs for Randomized Clinical Trial Participation 
Among Hispanics, 10 CANCER CONTROL84 (2003) (arguing Hispanic patients make decisions 
about participation based on the quality of the relationships they have with providers). 

66. For a discussion of the costs of care for patients enrolled in clinical trials, see Bruce 
H. Fireman et al., Cost of Care for Patients in Cancer Clinical Trials, 92 J. NAT'L CANCER 
INST. 136 (2000). 

67. Jane E. Brody, Ferreting for Facts in the Realm of Clinical Trials, N.Y. TIMEs, Oct. 
15,2002, atF7. 

68. See, e.g., Chan Ka Way, Acute Lymphoblastic Leukemia 32 CURRENT PROBS. 
PEDIA1RIC ADoU!SCBNT REALm CARE 40 (2002) •. 

69. Ftreman et al., supra note 66. 
70. Id. 
71. ld. 



2004] MANDATED COVERAGE OF EXPERIMENTAL THERAPY 231 

therapy may vary. Patients are also advised that their insurers may not cover 
the costs of participating in clinical trials. 72 

Indeed, insurers frequently deny coverage of therapy deemed "experi­
mental" or "investigational."73 Coverage can generally be negotiated for inter­
ventions that the patient would likely incur as part of disease management­
such as diagnostic studies needed to stage disease-whether or not the patient 
is in a clinical trial. In addition, coverage is generally provided for interven­
tions that are very similar to stl:}ndard therapy, such as modification of chemo­
therapy regimen. Also, coverage is typically provided for management of 
adverse reactions to experimental treatment. Expensive, novel interventions 
are most likely to be ones for which insurers have the incentive to delay 
decisions to allow coverage. A very high profile example of an intervention 
deemed experimental and hence, denied by many insurers was myoablative 
chemotherapy with stem cell support for breast cancer, where coverage was 
refused whether or not the patient was participating in a clinical trial. These 
refusals and legal challenges to them are discussed in Part five (V) below. 

When patients' health insurance does not cover participation in clinical 
trials, both individual and collective problems may arguably result. On the 
individual level, patients may incur unforeseen and non-reimbursed costs for 
health care. They may be denied care that might actually prove beneficial to 
them. Care studied in a clinical trial might also be offered off-study. Refusal 
by insurers to pay for care in the study might create incentives for patients to 
seek care off-study. The pattern thus created would be counterproductive both 
for patients and for health care generally. Patients may receive the care in a 
less structured setting in which they are not as well monitored. Consequently, 
because the care is not formally characterized as investigational, it may come 
to be regarded as established therapy even though there is not yet sufficient 
data to support that conclusion. However, because the care is being delivered 
off-study and not as research, clinicians will be deterred from collecting or 
publishing systematic data about side effects and outcomes. Collecting such 
data would require characterizing the delivery of the care as "research." 
Moreover, patients provided care off-study are not randomized, and apparently 
promising outcomes may be only anecdotal results of confounding factors 
rather than of the care itself. The upshot may be the worst of both worlds: 
patients receive care erroneously styled as "treatment" under circumstances 
where scientifically useful evaluative data is not and cannot be collected. 

In the fall of 2000, Medicare began covering routine costs for patients 
participating in specified clinical trials, as well as reasonable and necessary 

72. See, e.g., NCI, Cancer Facts, at http://cis.nci.nih.gov/fact/7 _ 41.htm Oast visited Feb. 
5, 2004) (answering questions for patients contemplating bone marrow transplants). 

73. See, e.g., NCI, CI.lNICAL TRIALs AND INSURANCE COVERAGE- A RESOURCE GUIDE: 

SUMMARY, at http:/lwww .nci.nih.gov/clinicaltrials!understanding/insurance-coverage Oast 
visited Mar. 16, 2004). 



232 INDIANA HEALm LAW REVIEW . [Vol. 1:213 

treatment for complications arising from participation in clinical trials.74 In 
extending this coverage, Medicare acted on the recognition that there was 
often little reliable data as to whether new therapies are safe or effective in the 
Medicare population. Extending coverage to patients in clinical trials meeting 
specified scientific criteria would help to remedy that gap, thus enhancing 
coverage decisions generally. For purposes of this coverage, "routine costs" 
were defined to include all of the items and services otherwise available to 
Medicare patients .. Thus, routine costs would include hospitalization or costs 
for administration. of drugs, but not the drugs themselves, to the extent that 
Medicare does not include drug coverage. Services specified as routine costs 
are conventional care, care required to administer the experimental therapy or 
prevent complications, and reasonable and necessary care for the diagnosis or 
treatment of complications. Routine costs exclude the costs of the test item 
(drug or medical device) itself. Also excluded are services provided solely for 
data collection and analysis rather than the patient's clinical needs, such as 
additional scans for monitoring disease progression and efficacy of the test 
item but not used for clinical management. Any item customarily provided 
free of charge for patients participating in the clinical trial (whether or not 
covered by Medicare) is also excluded as a routine cost. 

For care to be covered, the purpose of the trial must be an evaluation of 
a service that falls within a Medicare benefit category. Trials must have 
therapeutic intent and not be designed to test toxicity or disease patho­
physiology exclusively. Trials of therapeutic interventions must enroll only 
patients with a diagnosed disease. Healthy volunteers may be included as 
controls in tests of diagnostic interventions. Additional criteria that clinical 
trials qualifying for coverage must meet include: (1) the principal purpose of 
the trial must be to test whether the intervention potentially improves the 
participants' health outcomes; (2) the trial must be well-supported by available 
scientific and medical information or it must be intended to clarify or establish 
the health outcomes of interventions already in common clinical use; (3) the 
trial must not unjustifiably duplicate existing studies; (4) the trial design must 
be appropriate to answer the research question being asked in the trial; (5) the 
trial sponsor must be a credible organization or individual capable of execut­
ing the proposed trial successfully; (6) the trial must comply with federal 
regulations relating to the protection of human subjects; and (7) all aspects of 
the trial must be conducted according to the appropriate standards of scientific 
integrity. 

These criteria aim primarily to ensure the scientific merit of the study. 
The trial must be·designed to answer a research question, and duplication of 
results must be avoided. The criteria are also aimed at ensuring that the trial 

74 .. The following discussion can be attributed to the Medicare coverage document. See 
Ctrs. for Medicare & Medicaid Servs., National Coverage Determinations, at http://www .ems 
.hhs.gov/ncd/searchdisplay.asp?NCD_ID=l&NCD_ vrsn_num=l (last visited July 28, 2003). 



2004] MANDATED COVERAGE OF ExPERIMENTAL THERAPY 233 

meets core ethical standards for research: scientific integrity and the protec­
tion of human subjects. To be sure, the research must also have a therapeutic 
purpose; but therapeutic research that is not fitSt rate science wt11 not qualify 
for coverage under the Medicare standards. Clinical trials sponsored by the 
Nlll, the Centers for Disease Control and Prevention ("CDC"), the Agency for 
HealthCare Research and Quality ("AHRQ"), the· HealthCare Financing 
Administration ("HCFA," now ''CMS"), the Department ofDefense ("DOD"), 
and the Department of Veterans Affairs ("VA") are deemed to automatically 
meet these additional criteria. Trials conducted by centers Or cooperative 
groups, such as the Southwest Oncology Group ("SWOO"), supported by the 
listed federal entities also automatically meet the additional criteria. Trials 
conducted under an investigational new drug application reviewed by the FDA 
likewise automatically qualify. Other trials must certify·that they meet the 
Medicare criteria and are enrolled in a Medicare clinical trials registry. 
Further development and review of these criteria is assigned tO a multi-agency 
federal panel convened by the AHRQ. These procedures are designed to 
ensure scientific quality and research integrity without the need for study-by­
study review. 

Some insurers in the private market have adopted policies that parallel 
the approach taken by Medicare. Others continue to refuse to cover any 
therapy deemed experimental. State legislatures have entered the arena by 
enacting coverage mandates, but some of these mandates extend coverage far 
more broadly than the Medicare criteria. These issues are the subject of Part 
four(IV). 

N. STATIITORY APPROACHESTOCOVBRAGB 

Within the past ten years, many states have mandated insurance 
coverage for some care treated as experimental under insurance plans. The 
majority of states now have coverage mandates for some off-label drug use.75 

Fewer states mandate coverage for certain patients participating in clinical 

75. ALA. CODE§ 27-1-10.1 (2003); ARIZ. REV. STAT.§ 20-826(R) (2003); ARK. CODE 
ANN.§ 23-79-147(b) (Michie 2003); CAL. INS. CODE§ 10123.195(a)(2)(B) (Deering 2003); 
CONN. GEN. STAT. ANN.§ 38a-492b (West 2003); GA. CODE ANN.§ 33-24.59.11 (2003); 215 
fil...COMP.STAT. 5/370-r(2003);1ND.CODEANN. § 27-8-20-7 (Micbie2003);KAN.STAT.ANN. 
§ 40-2,168 (2002); LA. REv. STAT. ANN.§ 22:230.4(a) (West 2003); ME. REv. STAT. ANN. tit 
24, § 2320-F (2003); MD. CODE ANN., INS.§ 15-804 (2003); MAss. GEN. LAws ch. 175, § 47K 
(2003); MicH. COMP. LAWS§ 500,3406q (2003); MINN. STAT.§ 62Q.525 (2002); Miss. CODE 
ANN.§ 83-9-8 (2003); NEV.REV. STAT.§ 689A.0404 (2003); N.H. REv. STAT.ANN. § 415:6-
g(1)(2002);N.J.STAT.ANN. § 17B:26-2.1g(West2003);N.C.GEN.STAT. §58-51-59 (2003); 
N.D. CENT. CODE§ 26.1-36-06.1 (2003); OHIO REV. CODE ANN. § 1751.66(A) (1\n.derson 
2002); OKLA. STAT. tit 63, § 1-2604 (2002); OR. REv. STAT.§ 743.697(1) (2001); R.I. GEN. 
LAws § 27-55-2(a) (2002); S.C; CODE ANN. § 38-71-275(A) (Law. Co-op. 2002); S.D. 
CODIFIED LAWS§ 58-17-101 (Micbie2003); TENN. CODE ANN.§ 56-7-2352(2003);TEX.INS. 
CODE ANN. § 21.53M (Vernon 2003); VA. CODE ANN. § 38.2-3407.5(A) (Michie 2003). 



234 INDIANA HEALTH LAW REviEW [Vol. 1:213 

trials. 76 As state law governing insurance, these mandates are preempted by 
the Employee Retirement Income Security Act (''ERISA") for employee 
welfare plans not deemed to be insurance; ·thus, state mandates do not extend 
to employers who self-insure their health plans.77 Some employers nonethe­
less may follow state mandates voluntarily. As the descriptions below indi­
cate, however, the mandates vary in the scientific basis they require. The 
more indiscriminate mandates .arguably exacerbate the confusion between 
research and treatment. These mandates may lead patients to demand care that 
they. believe is needed treatment, even though it may not be beneficial, 
supported by evidence, or come under policy language covering "reasonable 
and necessary" care. 

A. Coverage Mandates for Off-Label Drug Use 

FDA approval of new drugs for marketing includes labeling with 
indication, dosage, and patient population. Labeling information is based on 
data supporting safety and efficacy of the drug for the indication in the patient 
population. 78 Once a drug is marketed, physicians are legally permitted to 
prescribe it for different patient populations, at different dosages, or for new 
indications. In such situations, the drug is used "off-label." Drugs used to 
treat cancer are frequently employed in off-label fashion.79 This practice is 
reasonable because cancer treatment evolves rapidly, and a drug initially 
found effective against one tumor type might also have promise against 
another. The alternative would be to require FDA approval for each new 
indication, a cumbersome and time-consuming process that would introduce 
substantial delays. On the other hand; indiscriminate off-label use risks 
subjecting patients to highly toxic therapy that has not undergone rigorous 

76. ARIZ. REV. STAT. ANN.§§ 20-826.01(A)(6),(7) (2003); CAL. INS. CODE§ 10145.4(a) 
(2003); CONN. GEN~ STAT. ANN. § 38a-504a (West 2003); GA. CODE ANN. § 33-24-59.l(a)(2) 
(2003); 2151IL. COMP.STAT.ANN. 5/356-y(c)(2) (2003); LA. REv. STAT. ANN.§ 22.230.4(E)(2) 
(West 2003); ME. REv. STAT. ANN. tit. 24-A, § 4310(l)(A) (2003); MD. CODE ANN., INS.§ 15-
827(e)(l)(2003);MAsS.OEN.LAWSANN.ch.175,§110L(c)(2),(4)(West2003);MICH.COMP. 
LAws§550.1929(7)(West2003);Mo.ANN.STAT.§376.429(1)(West2003);NEv.REv.STAT. 
ANN. 689A.04033 (Michie 2003); N.H. REV. STAT. ANN.§ 415:18-l(ll) (2003); N.M. STAT. 
ANN. § 59A-2243(A) (Michie 2003); N.C. GEN. STAT. § 58-3-255(a)(l) (2003); VT. STAT. 
ANN. tit. 8, § 4088b(a)(3) (2003); VA. CODE ANN. § 38.2-3418.8(D) (Michie 2003). 

77. 29 u.s.c .. § 1144 (2003). 
78. SeeFOOD&DRUGADMIN.,GUIDANCEFORlNDUSTRY:CUNICALSTUDIESSECTIONOF 

LABELING FOR PRESCRIPTION DRUGS AND BIOLOGics-CONTENT AND FORMAT (May 2001 
draft), available at http://www.fdagov/cder/guidance/1890dft.pdf 

79. Nat'l Cancer Inst., Understanding the Approval Process for New Cancer Drugs: Q&A: 
Off-Label Drugs, at www.cancer.gov/clinicaltrialslunderstandinglapproval-process-for-cancer­
drugslpage5 (last visited July 16, 2003). For a policy evaluating reimbwsement on the basis of 
each type of use, see, e.g., the Montana Blue Cross/Blue Shield web site, at http://mtl-bcbs. 
bcbsmt.comlbsd/MedicalPolicy .nsf/PoliciesByTitleView?Open View&Count=300 (last visited Feb. 
7,2004). 



2004] MANDATED COVERAGE OF EXPERIMENTAL THERAPY 235 

scientific evaluation. Thus, the NCI both recognizes and urges caution in off­
label use. 80 It is estimated that at least half of all cancer patients receive some 
off-label treatment.81 Estimates of off-label use are higher for children, 
because drugs are frequently approved for adults without having been tested 
in children. 82 

More than thirty states currently mandate at least some coverage for off­
label uses of drugs.83 Some states apply the mandate to off-label use in cancer 
treatment only, 84 while others apply the mandate generally. 85 Still others apply 
the mandate only to drugs used in treating life-threatening or chronic and 
seriously debilitating conditions. 86 These mandates typically leave other 
coverage limitations undisturbed, such as deductibles and co-payments or 
restrictions on the use of non-formulary drugs. 87 

80. Nat'l Cancer Inst., supra note 79. 
81. T. Laetz & G. Silberman, Reimbursement Policies Constrain the Practice of 

Oncology, 266 JAMA 2996 (1991). 
82. Sharon Conroy et al., Unlicensed and Off lAbel Drug Use in Acute Lymphoblastic 

Leukaemia and Other Malignancies in Children, 14 ANNALS ONCOLOGY 42 (2003). 
83. Nat'l Cancer Inst., Understanding the Approval Process for New Cancer Drugs: 

Summary,atwww.cancer.gov/clinicaltrials/understandinglapproval-process-for-cancer-drugs/ 
pageS#anchor-wha-15227 (last visited July 18, 2003) (indicating that thirty-nine states have 
such coverage requirements, by statute, administrative regulation, or non-legislative agreement). 
However, this count may be an overestimation. Several listed states provide for review of 
coverage denials but do not appear to mandate coverage for off-label drugs. One listed state, 
Kentucky, has repealed its mandate. KY. REv. STAT. ANN.§ 304.17A-137 (repealed 2002). 
Also, Louisiana covers off-label uses only within clinical trials. LA. REv. STAT. ANN. 
§ 22:230.4(a) (West 2003). Thirty states currently have statutory mandates for coverage of at 
least some off label uses. See supra note 75. 

84. ARiz. REV. STAT.§ 20-826(R) (2003); ARK. CODE ANN.§ 23-79-147(b) (Michie 
2003); CONN. GEN. STAT. ANN.§ 38a-492b (West 2003); 2151I.L. CoMP. STAT. 5/370-r (2003); 
KAN. STAT. ANN. § 40-2,168 (2002); ME. REv. STAT. ANN. tit. 24, § 2320-F (West 2003); 
MINN. STAT. § 62Q.525 (2002); MISS. CODE ANN. § 83·9-8 (2003); NEV. REV. STAT. 
§ 689A.0404(2003);N.C.GEN. STAT.§ 58-51-59 (2003); OKlA. STAT. tit. 63, § 1-2604(2002); 
R.I. GEN. LAWS§ 27-55-2(a) (2002); S.C. CODE ANN.§ 38-71-275(A) (Law. Co-op. 2002); 
TENN. CODE ANN.§ 56-7-2352 (2003); VA. CODE ANN.§ 38.2-3407.5(A) (Michie 2003). 

85. AU.. CODE§ 27-1-10.1 (2003); IND. CODE ANN.§ 27-8-20-7 (Michie 2003); MD. 
CODE ANN., INS. § 15-804 (2003); MASS. GEN. LAws ch. 175, § 47K (2003); N.H. REV. STAT. 
ANN.§ 415:6-g(l) (2002); N.J. STAT. ANN.§ 17B:26-2.1g (West 2003); N.D. CENT. CODE, 
§ 26.1-36-06.1 (2003); OmoREv.CODEANN. § 1751.66(A) (Anderson2002); OR. REV. STAT. 
§ 743.697(1) (2001). 

86. CAL. INS. CODE § 10123.195(a)(2)(B) (l>eering 2003); MICH. COMP. LAWS § 
500.3406q (2003); TEX. INS. CODE ANN. § 21.53M (Vernon 2003). Cf. S.D. CODIFIED LAWS 
§ 58-17-101 (Michie 2003) (including limit only to cancer and life-threatening conditions). 

87. ARlZ. REV. STAT. § 20-826(R) (2003); CAL. INS. CODE § 10123.195(h) (Deering 
2003); ME. REv. STAT. ANN. tit. 24, § 2320-F(2)(E) (West 2003); N.H. REv. STAT. ANN. § 
415:6-g(3) (2002); OHIO REV. CODE ANN.§ 1751.66(0)(4) (Anderson 2002); R.I. GEN. LAWS 
§ 27-55-2(d) (2002). 



236 INDIANA HEALTH LAW REVIEW [Vol. 1:213 

Many states require at least some verification that use of the drug has 
been scientifically validated. 88 Recognized validation methods include drug 
compendia such as the United States Pharmacopeia Drug Information, the 
American Medical Association Drug Evaluations, or the American Hospital 
Formulary Service Drug Information; or publications in the peer-reviewed 
literature. 89 These statutes provide patients with assurance that coverage will 
not be denied for beneficial treatment merely because sufficient evidence of 
safety and efficacy has not been provided to the FDA to authorize changes in 
the drug's labeling. To the extent validation in the scientific literature is 
required, coverage is limited to therapy that has been evaluated for safety and 
efficacy. On the other hand, if the quality of scientific evidence is low, the 
mandates open the door to coverage for therapy of uncertain efficacy. 

The quality of scientific evidence required in the mandates varies 
significantly. Some states require only a single peer-reviewed article, without 
further discussion of the required strength of the scientific evidence. In 
theory, the article could be a case report of the drug's use in one patient or in 

88. The Arizona statute has perhaps the most complete explanation of what might be 
considered scientific validation. In addition to medical reference compendia, Arizona will 
recognize data if: 

(a) At least two articles from major peer reviewed professional medical journals 
have recognized, based on scientific or medical criteria, the drug's safety and 
effectiveness for treatment of the indication for which the drug has been pre­
scribed; (b) No article from a major peer reviewed professional medical journal 
has concluded, based on scientific or medical criteria, that the drug is unsafe or 
ineffective or that the drug's safety and effectiveness cannot be determined for 
the treatment of the indication for which the drug has been prescribed; 
[and] (c) The literature meets the uniform requirements for manuscripts sub­
mitted to biomedical journals established by the international committee of 
medical journal editors or is published in a journal specified by the United States 
department of health and human services as acceptable peer reviewed medical 
literature pursuant to section 186(t)(2)(B) of the social security act, 42 U.S.C. § 
186(t)(2)(B) (2003). 

ARIZ. REv. STAT. § 20-826(s)(2) (2003). 
Other states require validation in at least two peer-reviewed articles. Miss. CODE ANN. 

§ 83-9-8 (2003); NEV. REv. STAT. 689A.0404 (2003); S.C. CODE ANN. § 38-71-275(C)(3) 
(Law. Co-op. 2002). Some states, however, require as little as validation in one peer-reviewed 
article. IND. CODE ANN.§ 27-8-20-7 (Michie 2003); N.J. REV. STAT.§ l7B:26-2.1g (2003). 
Yet others refer generally to the "medical literature," without specifying numbers of articles or 
size of studies. MASS. ANN. LAWS ch. 175, § 47K (Law. Co-op. 2003); TENN. CODE ANN. § 56-
7-2352(c)(1) (2003). Oregon requires the "majority" of the medical literature. OR. REv. STAT. 
§ 743.697(1)(b) (2001). Virginia notes specifically that "peer-reviewedmedicalliterature does 
not include publications or supplements to publications that are sponsored to a significant extent 
by a pharmaceutical manufacturing company or health carrier." VA. CODE ANN. § 38.2-
3407.5(C) (Michie 2003). Even in the peer-reviewed literature, studies may vary importantly 
in the quality of the evidence they provide for a given therapeutic intervention. Some 
publications may consist of single case reports or case series with historical controls, thought 
to be evidence of lesser quality than randomized clinical trials. See Medicare Coverage 
Advisory Comm., supra note 7. 

89. See, e.g., CONN. GEN. STAT. ANN. § 38a-492b (West 2003). 



2004] MANDATED COVERAGE OF ExPERIMENTAL THERAPY 237 

a small series of patients. Moreover, blanket coverage mandates do not 
encourage systematic collection of further scientific data about the efficacy of 
therapy, although rigorous studies may take place in the clinical context. 

Two solutions to these difficulties might be suggested. The first is to 
limit these coverage mandates to off-label uses that have reached sufficient 
scientific validation to appear in one of the recognized pharmaceutical com­
pendia. A second possibility, designed especially to gather good scientific 
evidence of therapeutic safety and efficacy, would be to cover off-label uses 
in appropriately-designed clinical trials, as illustrated by many of the clinical 
trials involving children with cancer. 

B. Coverage Mandates for Participation in Clinical Trials 

Since 1998, seventeen states have enacted statutes extending coverage 
mandates for cancer patients in at least some clinical trials. 90 Despite the 
growth of such statutes, only one law review has regarded the development as 
a "note" -worthy innovation in state law. 91 Even law reviews in Connecticut, 
the home of much of the insurance industry, left the development un­
remarked. One early article, published before the movement towards such 
statutes gathered full steam, argued for limited coverage of patients participat­
ing in trials meeting certain standards of clinical integrity, scientific integrity, 
and fiscal soundness.92 This discussion saw the primary purpose of such 
coverage as a response to patients' desires for last chance therapy, and the 
secondary purpose as the development of better data about new therapies. 

The statutory coverage mandates actually enacted vary in important 
respects with regard to patients covered, type of clinical trial, costs covered, 
mechanism of operation, and apparent rationale. Georgia's statute, the very 
first, only covers participation in clinical trials for children diagnosed with 
cancer.93 Most other statutes are limited to patients who have been diagnosed 
with cancer and meet the qualifications for participating in the trial.94 Nevada 
has also extended the mandate to patients enrolled in clinical trials for 
treatment of chronic fatigue syndrome.95 Mandates in Maine, Maryland, New 
Hampshire, and North Carolina extend even more broadly ·to patients in 
clinical trials for any life-threatening condition. 96 

90. See statutes cited, supra note 76. 
91. Lesley C. Murphy, Health and Welfare: Mandatory Health Insurance Coverage for 

Cancer Clinical Trials, 33 MCGEORGEL. REv. 314,314-22 (2002). 
92. Mary Ader, Investigational Treatments: Coverage, Controversy, and Consensus, 5 

ANN. HEALTHL. 45,45-60 (1996). 
93. GA. CODE ANN. § 33-24-59.1(a)(2) (2003). 
94. See, e.g., CAL. INS. CODE§ 10145.4(a) (West 2003). 
95. NEV. REV. STAT. ANN. 689A04033 (Michie 2003). 
96. See ME. REv. STAT. ANN. tit. 24-A, § 4310(1)(A) (2003); Mo. CODE ANN., INs.§ 15-

827(e)(1)(ii) (2003); N.H. REv. STAT. ANN. § 415:18-l(II) (2003); N.C. GEN. STAT. § 58-3-
255(a)(l) (2003). 



238 INDIANA HEALTH LAW REVIEW [Vol. 1:213 

States have different limits on the phases of clinical trials covered. 
Covering all phases (Phases I-N) are California, Maryland, Massachusetts 
(Phase I only at an academic medical center or affiliated facility), New 
Hampshire (Phase I and ll on a case-by-case basis only), Vermont (all Phases 
limited to approved centers), and Virginia (Phase I on a case-by-case basis 
only).97 lllinois, Louisiana, Nevada, and New Mexico only cover Phases ll­
N.98 Connecticut only covers participation in Phase ill trials,99 while 
Missouri covers Phases ill and N.100 Statutes covering Phase I trials require 
insurers to pay for interventions that do not necessarily have the primary goal 
of treating the patient at hand. Statutes limiting coverage to the later phases 
focus more directly on trials in which therapeutic goals are paramount. 

Many, but not all, of these statutes also limit the mandate explicitly to 
trials with some anticipated benefit for the patient, although they formulate 
this requirement in different ways. Some statutes require a determination that 
participation offers the possibility of clinical benefit to the patient. 101 Other 
statutes specify that qualifying trials must have therapeutic intent and end 
points not defined exclusively to test toxicity.102 Others require a reason to 
believe that the proposed treatment is at least as effective as, or better than, 
available treatment.103 Louisiana covers trials with therapeutic or palliative 
intent for prevention or early detection of cancer. 104 In a statute in effect from 
2000-2003, lllinois required that the care be both medically appropriate and 
that it not yet have been determined to be safe, effective, or otherwise 
medically appropriate or necessary.105 Perhaps reflecting an error in drafting, 
Missouri requires that "there must be equal to or superior, noninvestigational 
treatment alternatives and the available clinical or preclinical data must 
provide a reasonable expectation that the treatment will be superior to the 

97. CAL INS. CODE § 10145.4(a) (West 2003); MD. CODE ANN., INS. § 15-827(e)(1) 
(2003); MAss. GEN. LAWS ANN. ch. 175, § 110L(c)(2), (4) (2003); N.H. REV. STAT. ANN.§ 
415:18-1(11) (2003); VT. STAT. ANN. tit. 8, § 4088b(a)(3) (2003); VA. CODE ANN.§ 38.2-
3418.8(D) (Michie 2003). 

98. 215 IlL COMP. STAT. ANN. 5/356-y(c)(2) (West 2003); LA. REV. STAT. ANN. § 
22.230.4(E)(2) (West 2003); NEV. REV. STAT. ANN. 689A04033 (Michie 2003); N.M. STAT. 
ANN. § 59A-22-43(A) (Michie 2003); N.C. GEN. STAT. § 58-3-255(a)(1) (2003). 

99. CoNN. GEN. STAT. ANN.§ 38a-504a (West 2003). Perhaps the assumption is that 
Phase IV, post-marketing trials, will be covered in any event, so their addition is not necessary. 

100. Mo. ANN. STAT. § 376.429(1) (West 2003). 
101. CALINS.CODE§ 10145.4(a)(West2003)(insured'sphysiciandetermines"meaning­

ful potential to benefit"). 
102. Id.; MAss.GEN.LAWSANN. ch.175, § 110L(c)(9)(West2003)(addsthattheremust 

be a reasonable expectation of a medical benefit commensurate with the risks of participation). 
103. ARiz. REv. STAT. ANN. §§ 20-826.01(A)(6)-(7) (West 2003); MD. CODE ANN., INS. 

§§ 15-827(e)(4)-(5) (2003); NEV. REv. STAT. ANN. 689A04033.1(d)-(e) (Michie 2003); N.H. 
REv. STAT. ANN. § ill(b) (2003); VA. CODE ANN. §§ 33.2-3418.8(G)(l)-(2) (2003). If 
investigational treatment did not meet this standard, would the clinical trial be ethical? 

104. LA. REV. STAT. ANN. §§ 22:230.4(E)(1)-(2) (West 2003). 
105. 2151IL. COMP. STAT. ANN.§§ 5/356y(a), (c)(3) (repealed 01/0112003). Illinois also 

limited the coverage mandate to $10,000 annually. Id. § 5/356y(a). 



2004] MANDATED COVERAGE OF EXPERIMENTAL THERAPY 239 

noninvestigational alternatives." 106 More clearly, Maine requires "meaningful 
potential for significant clinical benefit."107 North Carolina requires that 
participation in the trial be medically indicated and more preferable for the 
patient than standard care, although it is difficult to see how a trial would be 
ethical if participation in it were medically indicated because one of the 
options is already known to be better than standard care. 108 

As additional protection for patients, some states' statutes add 
provisions about the qualifications of the physicians providing the care. 
Arizona, Louisiana, Nevada, New Hampshire, New Mexico, and Virginia all 
require that the care be within the scope of practice, experience, and training 
of the physicians involved. 109 Massachusetts, Missouri, and North Carolina 
add that the physicians must treat a sufficient volume of patiepts in that area 
of medicine to maintain their expertise.uo illinois' repealed statute required 
that the patient's primary care physician, if any, be involved in coordinating 
the care. 111 Massachusetts and New Hampshire require that the patient give 
informed consent that meets current legal and ethical standards, 112 and Nevada 
demands that the patient be informed of the procedure, alternative methods of 
treatment, and the risks associated with participation in the clinical trial. 113 

In addition to implementing medical provider guidelines, states employ 
a number of methods to ensure that covered trials meet scientific standards. 
The ones of these that are most stringent parallel Medicare's limit to trials 
approved by various federal agencies or the centers and cooperative groups 
they support. 114 Maine limits the approving federal agencies to the Depart­
ment of Health and Human Services, the Nlli, and Nlli-approved cooperative 
groups or centers. 115 Arizona adds trials approved by a panel of experts in 

106. Mo. ANN. STAT. § 376.429(2) (West 2003). 
107. ME. REv. STAT. ANN. tit. 24-A, § 4310(l)(c) (2003). 
108. N.C. GEN. STAT. § 58-3-255(a)(l)(a) (2003). The exact language in the North 

Carolina statute reads: "medically indicated and preferable for that patient compared to 
available non-investigational treatment alternatives and ... have clinical and preclinical data 
that show the trial will likely be more effective for that patient than available noninvestigational 
alternatives." !d. 

109. ARIZ. REV. STAT. ANN. § 20-826.01(A)(5) (West 2003); LA. REv. STAT. ANN. 
§ 22:230.4(E)(5) (West 2003); NEV. REv. STAT. ANN. 689A.04033.1(c) (Michie 2003); N.H. 
REv. STAT. ANN. § 415:18- 1(ill)(c) (2003); N.M. STAT. ANN. § 59A-22-43(A)(7) (Michie 
2003); VA. CODE ANN. § 38.2-3418.8(F) (Michie 2003). 

110. MASS. GEN.LAWSANN. ch. 175, § 110L(c)(3)(2003); Mo. ANN. STAT.§ 376.429(2) 
(2003); N.C. GEN. STAT. § 58-3-255(a)(1)(c) (2003). 

111. 215 ILL. COMP. STAT. ANN. § 5/356-y(c)(5) (repealed 2003). 
112. MASS. GEN. LAWS ANN. ch. 175, § 110L(c)(6) (West 2003); N.H. REv. STAT. ANN. 

§ 415:18-l(Vl) (2003). 
113. NEV. REv. STAT. ANN. 689A.04033.1(g) (Michie 2003). 
114. CAL. INS. CODE § 10145.4(c) (West 2003) (including trials of uses of non­

experimental drugs); CONN. GEN. STAT. § 38a-504b (2003); 215 ILL. COMP. STAT. ANN. § 
5/356-y(c)(3) (repealed 2003); LA. REv. STAT. ANN.§ 22:230.4(E)(3) (West 2003); NEV. REV. 
STAT. ANN. 689A.04033, l(b) (Michie 2003). 

115. ME. REv. STAT. ANN. tit. 24-A, § 4310(4) (West 2003). 



240 INDIANA HEALTH LAW REVIEW [Vol. 1:213 

clinical research within academic health institutions in Arizona.u6 Maryland, 
Massachusetts, Missouri, New Hampshire, and Virginia add trials approved 
by institutional review boards of institutions with multiple project assurance 
agreements with the Office of Protection from Research Risks of the NCI. 117 

Many states add the requirement that the trial receive institutional review 
board approval, required for protection of human subjects in research, despite 
the fact that this is already a requirement for any federally funded research. 118 

Like Medicare, these state statutes typically mandate coverage for 
routine costs of patient care: services provided outside a clinical trial as 
standard care, services provided for drug administration, services provided for 
disease staging and appropriate clinical monitoring, and services provided to 
treat side-effects and complications.119 Costs typically not covered are those 
that are associated with data collection and management, 120 and costs that 
arguably should be covered if the goal of the statute were to advance medical 
knowledge. Another typical exclusion is non-clinical services, such as travel, 
housing, and companion expenses. 121 Unapproved drugs or devices are not 
covered in Arizona and California. 122 Another common exclusion is any costs 
covered by the trial for other enrolled patients, so insurers are not billed for 
costs the trial is willing to fund for other patients. 123 Some of the statutes limit 
costs to in-state care, 124 others to in-state care unless the care is only available 
out-of-state125 or out-of-state care is otherwise a covered benefit for standard 

116. ARiz. REV. STAT. §§ 20-826.0l(A)(3)(f)-(g) (2003). 
117. Mo. CODE ANN., INs. § 15-827(e)(2)(v) (2003); MASS. GEN. LAWS ANN. ch. 175, 

§§ llOL (c)(2), (4) (West 2003) (requiring that Phase I trials be conducted at an academic 
medical center or affiliated facility by clinicians with staff privileges); Mo. REV. STAT. § 
376.429.4(5) (2003) (statutory language requiring "appropriate assurance" rather than "multiple 
project assurance"); N.H. REV. STAT. ANN. § 415:18-l(ill)(a)(5) (2003); VA. CODE ANN. § 
38.2-3418.8(E)(5) (Michie 2003). 

118. ARIZ. REV. STAT.§ 20-826.0l(A)(4) (2003); LA. REv. STAT. ANN.§ 22:230.4(E)(4) 
(West 2003); N.M. STAT. ANN. § 59A-22-43(A)(6) (Michie 2003). 

119. CAL. INS. CODE§ 10145.4(b)(l) (West 2003). 
120. ARIZ. REv. STAT.§ 20-826.01(G)(5)(d) (2003); CAL. INS. CODE§ 10145.4(b)(2)(C) 

(West 2003); LA. REv. STAT. ANN. § 22:230.4(A)(10)(b) (West 2003); ME. REV. STAT. ANN. 
tit. 24-A, § 4310(2) (West 2003); MD. CODE ANN., INS.§ 15-827(a)(7)(ii)(3) (2003); MASS. 
GEN. LAWS ANN. ch. 175, § 110L(a)(7) (West 2003); N.H. REV. STAT. ANN. § 415:18-
l(l)(h)(4) (2003); NEV. REV. STAT. ANN. 689A.04033.4(g) (Michie 2003); N.C. GEN. STAT.§ 
58-3-255(d) (2003); VA. CODE ANN.§ 38.2-3418.8(C) (Michie 2003). 

121. ARIZ. REv. STAT.§ 20-826.01(G)(5)(c) (2003); CAL. INs. CODE§ l0145.4(b)(2)(B) 
(West 2003); LA. REv. STAT. ANN.§ 22:230.4(A)(10)(a) (West 2003); NEV. REV. STAT. ANN. 
689A .04033.4(e)-(f) (Michie 2003); N.C. GEN. STAT. § 58-3-255(d) (2003). 

122. ARIZ. REv. STAT. § 20-826.01(G)(5) (a)-(b) (West 2003) (excluding coverage on 
drugs or devices in Phase I trials and any investigational drugs or devices); CAL. INS. CODE 
§ 10145.4(b)(2)(A) (West 2003). 

123. CAL. INS. CODE§ 10145.4(b)(2)(E) (West 2003). 
124. ARiz. REV. STAT. § 20-826.0l(G)(5)(f) (2003); NEV. REV. STAT. § 689A.0404.1(f) 

(2003). 
125. CAL. INS. CODE§ 10145.4(d) (West 2003). 



2004] MANDATED COVERAGE OF EXPERIMENTAL THERAPY 241 

treatment in the patient's insurance policy. 126 Louisiana, Maryland, and 
Missouri have the additional requirement that any entity seeking coverage for 
clinical trials must maintain an up-to-date electronic post of the approved 
clinical trials, including the phase, approving entity, covered cancers, and 
estimated number of participants. 127 

Most of the statutes directly mandate benefits. However, the mandates 
may contain important exclusions. California, for example, excludes Medi­
care supplementary insurance and CHAMPUS, as well as insurance policies 
limited to certain types of benefits (for example, disability, vision only). 128 

Louisiana excludes plans paid for under Title XVII (SSI for the aged and 
disabled) or Title X (SSI for the blind) of the Social Security Act.129 Other 
statutes are structured as limits on what may be required as reimbursement for 
patients in clinical trials of cancer treatment; Arizona and New Mexico 
statutes function this way, and state explicitly that they provide only an 
administrative remedy, not a private right of action brought by a patient 
against an insurance corporation. 130 The Michigan statute creates a fund to 
defray costs of patient participation, up to $5000 per facility in Phase m or IV 
studies sponsored by the NIH; 131 revenues for the fund come from fmes paid 
for violations of the insurance code, a strategy that avoids ERISA preemption 
under the rationale of Blue Cross & Blue Shield v. Travelers .132 

Most importantly, statutes differ in apparent rationale. No statute sets 
out a legislative statement of purpose, so the reasoning behind the statutes 
must be gleaned from their structure and substance. From the account given 
above, it would appear that the principal aim is likely therapeutic value, or at 
least the minimum requirement that the trial be likely to provide the patient 
with prospects at least as good as those offered by standard therapy. Cost may 
have played a secondary role in the design of some statutes; for example, 
Vermont regards its mandate as a pilot program at approved medical centers 
and requires an analysis of cost data to determine the financial effect on 
insurance premiums. 133 Deference to approval by federal agencies would 
seem to assure scientific merit, but this approach may also have been moti­
vated by the effort to assure patients the best therapeutic options. Limited 
attention has been given to other scientific goals. lllinois wanted covered 
trials to be conducted for the primary purpose of determining safety and 

126. N.M. STAT. ANN.§ 59A-22-43(G) (Michie 2003). 
127. LA. REV. STAT. ANN. § 22:230.4(F) (West 2003); MD. CODE ANN., INS. § 15-

827(g)(1) (2003); Mo. REV. STAT.§ 376.429(5) (2003). 
128. CAL. INS. CODE § 10145.4(f) (West 2003). 
129. LA. REV. STAT. ANN. § 22:230.4(C) (West 2003). 
130. ARIZ. REv. STAT.§ 20-826.0l(D) (2003); N.M. STAT. ANN.§ 594-22-43(E) (Michie 

2003). 
131. MICH. COMP. LAWS § 550.1929(7) (2003). 
132. N.Y. State Conference of Blue Cross & Blue Shield Plans v. Travelers Ins. Co., 514 

u.s. 645 (1995). 
133. VT. STAT. ANN. tit. 8, § 4088b(a)(4) (2003). 



242 INDIANA HEALTH LAW REVIEW [Vol. 1:213 

efficacy, conducted at multiple centers, and aimed at the submission of peer 
reviewed publications.134 Massachusetts incorporates the Medicare require­
ment that the trial not unnecessarily duplicate existing studies, a criterion 
focusing on the scientific importance rather than the therapeutic efficacy of 
the trial. 135 The differentiation between trial phases for coverage purposes 
may also convey the message that not all clinical trials are equally therapeutic 
in intent. 

Otherwise, statutes leave the general impression that their legislative 
aim is to provide patients with the therapeutic benefits of clinical trials. This 
justification may leave patients who do not fit the entry criteria for the trials 
feeling that they have been dealt with unjustly because they are not receiving 
treatment that the state apparently believes is beneficial for others. Patients 
who do not meet strict entry criteria based on the science of the trials or who 
would like to participate in trials not covered by the statutes may wonder why 
they cannot take advantage of the therapy offered. Indeed, the one published 
commentary on these statutes, the assessment of the California statute in the 
McGeorge Law Review, illustrates this exact confusion about the rationale for 
the California mandate. 136 On the one hand, the commentary praises the 
statute for the possibility that by encouraging participation in clinical trials it 
will help to develop better information about cancer treatment. 137 On the other 
hand, it voices the concern that the statutory mandate will not cover therapy 
for a sufficient range of patients. 138 The commentary criticizes the require­
ment that the clinical trial must have "meaningful potential" to benefit the 
patient as vague and likely to exclude too many patients and praises the statute 
for increasing the likelihood that desperate patients (such as a woman with 
metastatic breast cancer seeking high-dose chemotherapy) will have access to 
coverage.139 

The limits in most statutes to trials approved by federal agencies will 
probably keep the actual coverage mandates within the bounds of rigorous 
scientific studies. Not all states impose these limits, however. It is important 
for legislatures to recognize that therapy alone cannot be the principal 
rationale for coverage, especially for patients participating in Phase I trials. 
Otherwise, statutory policy may contribute to the confusion held by many 
patients and members of the public that new therapies under investigation are 
the latest in established, beneficial care. This confusion has been especially 
apparent in the courts. 

134. 215lLL. COMP. STAT.§§ 5/356y(c)(3), (4), (6) (repealed 2003). 
135. MAss. GEN. LAws ch. 175, § 110L(c)(8) (2003). 
136. Murphy, supra note 91. 
137. /d. at 321. 
138. /d. at 322. 
139. /d. at 321. 



2004] MANDATED COVERAGE OF ExPERIMENTAL THERAPY 243 

V. HIGH-DOSE CHEMOTHERAPY IN THE COURTS 

The confusion between investigational therapies and beneficial treat­
ment has been manifested in the lawsuits brought by patients with breast 
cancer against insurance companies for denying coverage of high-dose chemo­
therapy. These lawsuits emerged in 1989, as initial encouraging reports 
appeared about the high-dose procedure and physicians began referring their 
patients to centers performing it. Patients were successful in many cases in 
gaining preliminary injunctions ordering coverage of the procedure and sub­
sequently, summary judgment in their favor. Only when insurance contracts 
listed the procedure explicitly as non-covered, and sometimes not even then, 
did the courts uphold coverage denials. 

Almost all of the cases denying coverage involved a determination of 
whether the patient could reasonably expect the care to be covered under their 
insurance contracts.14° Contracts create expectations, and courts interpreting 
contracts may have expectations about how similar contractual language is to 
be understood in other cases. In many of these cases, courts also determined 
that there had been sufficient testimony to support the efficacy of the therapy 
for the patient. Decisions about the efficacy of therapy, for the patient before 
the court, should have been case-specific, but as more of the cases were 
decided, courts cited prior holdings to support the conclusion that the therapy 
was beneficial to patients, thus multiplying the decisions' effects. The fre­
quency of reported decisions diminished in the late 1990s, after reports 
questioning the evidentiary support for high-dose chemotherapy for breast 
cancer began to appear in the literature. Nevertheless, patients continued to 
experience success in litigation, with rulings as late as 2002 holding that it 
was arbitrary and capricious for insurers to deny the coverage as investiga-

140. A few of the cases were state law contract cases. The six reported state cases resolv­
ing the issue on the merits (five mandating the coverage, one as late as 2002) occurred in 
Arkansas, Colorado, Illinois, and Michigan. Am. Inv. Life Ins. Co. v. Butler, 65 S.W.3d 472 
(Ark. Ct. App. 2002); Tepe v. Rocky Mtn. Health & Med. Serv., 893 P.2d 1323 (Colo. Ct. App. 
1994); O'Rourke v. Access Health, Inc., 668 N.E.2d 214 (Ill. App. Ct. 1996) (judgment for the 
insurer on plan language referring specifically to whether the federal Office of Technology 
Assessment had recommended coverage); Lubeznik v. HealthChicago, Inc., 644 N .E.2d 777 (Ill. 
App. Ct. 1994); Mich. Affiliated Healthcare Sys., Inc. v. Peoples Life Ins. Co., No. 225067, 
2001 Mich. App. LEXIS 1805 (Mich. Ct. App. Nov. 27, 2001); Taylor v. Blue Cross-Blue 
Shield of Mich., 517 N .W.2d 864 (Mich. Ct. App. 1994). An additional, recent state court deci­
sion resulted in a remand to the trial court for a decision on the merits. Pacino v. Med. Mut. of 
Ohio, Inc., No. 78757,2001 Ohio App. LEXIS 5749, at *8-9 (Ohio Ct. App. Dec. 20, 2001) (re­
jecting trial court's finding that contract language was ambiguous; remanding for trial on the 
merits). 

By far, the majority of the cases were either brought in federal court or removed to 
federal court because they involved coverage determinations under ERISA plans. In the cases 
under ERISA, contract interpretation is a question of law for the court. Coverage denials are 
reviewed under an arbitrary and capricious standard if discretion is reserved to the plan 
administrator; nearly all contemporary health insurance contracts reserve discretion in this way. 
Whether a contract provision is ambiguous is a question of law for the court; ambiguities are 
resolved against the insurer that has interests in denying the care. 



244 INDIANA HEALTH LAW REVIEW [Vol. 1:213 

tional. These rulings have appeared in many circuits, no doubt encouraging 
patients in the twin beliefs about their contractual coverage and the benefits 
of care. 141 

In reaching their conclusions, courts have been signally impressed by the 
testimony of physicians recommending the care. Courts reason that if physi­
cians recommend the care, it must be beneficial for their patients; therefore, 
it is irrational for insurers to deny coverage under an exception for care that 
is still under investigation and has not yet been demonstrated to be beneficial. 
In ruling against insurance plans, courts were especially impressed by their 
economic conflicts of interest, but almost never considered the possibility that 
physicians offering the procedure and testifying in its favor might have similar 
conflicts. 142 

The leading case ordering coverage ofHDC/ABMT, Dozsa v. Crum & 
Forster Insurance Co., was brought in 1989 by a patient with multiple 
myeloma. 143 In that case, the plaintiff was granted a preliminary injunction 
ordering the therapy. 144 To grant the injunction, the court was required to find 
that the plaintiff was likely to prevail on the merits, that failure to grant the 
relief would result in irreparable injury, that the insurer had not shown 
countervailing injury, and that the public interest was benefited by the grant 
of the relief. 145 Like subsequent courts relying on this decision, the court had 
little difficulty in finding that these standards were met. The language in the 
plaintiffs insurance contract stipulated that coverage would not be allowed 
unless the therapy was "neither educational nor experimental in nature nor 
provided primarily for research purposes."146 Prudential, the insurer, con­
tended that: 

A procedure or course of treatment does not become accepted 
in the profession until it has been demonstrated through clini­
cal trials as being safe and effective for the condition. One 
of the best ways to determine the acceptance of such a 
procedure is through the published medical literature. Cover­
age usually commences after there is clear consensus in peer 
reviewed medical literature. 147 

141. For a description of how early decisions encouraged patients to go to court to seek 
coverage and how insurers responded by changing contract language, see Goepel v. Mail 
Handlers Benefit Plan, No. 93-3711, 1993 U.S. Dist. LEXIS 13346 (D.N.J. Sept. 23, 1993), 
vacated by 36 F.3d 306 (3d Cir. 1994). 

142. See, e.g., Dozsa v. Crum & Forster Ins. Co., 716 F. Supp. 131, 139 (D.N.J. 1989). 
143. ld. 
144. ld. 
145. ld. at 137. 
146. ld. at 134 (emphasis in original). 
147. ld. at 135. Prudential had recently reviewed HDC/ ABMT for myeloma and concluded 

it did not meet this standard because it was still "investigational." Dozsa, 716 F. Supp. at 135. 
At the time, there were two published papers in the literature, both published by a single 
physician, Bart Barlogie; no randomized clinical trials had been reported. ld. at 136. Such case 



2004] MANDATED COVERAGE OF EXPERlMENTAL THERAPY 245 

The court concluded that Prudential was applying a new standard for "experi­
mental," which the court labeled "investigational," in reasoning this way. The 
plaintiff's treating oncologist had referred him to Johns Hopkins, where he 
was found to be a suitable candidate for the therapy. The court reasoned that 
if respected physicians were recommending the treatment, it must not be 
"experimental" within the medical community: 

While it is true that a lack of peer reviewed medical literature 
might constitute some evidence of nonrecognition of the 
therapy throughout a doctor's profession, it is certainly not 
conclusive evidence. The four very talented expert witnesses 
in the present case were in disagreement on the question 
whether the literature concerning ABMT treatment for multi­
ple myeloma constitutes backing of the treatment by a con­
sensus in peer reviewed medical literature and thus suppor­
tive of coverage even under defendants' mistaken reading of 
the Plan language. It is unnecessary to decide that question 
one way or another because the undisputed testimony of Dr. 
Wingard and Dr. Davis establishes 1) as proposed for plain­
tiff the treatment was not educational or experimental in 
nature and was not to be provided primarily for research pur­
poses and 2) it is commonly and customarily recognized 
throughout Dr. Wingard's profession as appropriate treatment 
of multiple myeloma. 148 

Many subsequent decisions involving patients with breast cancer likewise 
credited the testimony of treating oncologists over appeals to evidence-based 
medicine, as described in what follows. 

The first reported decision ordering coverage of HDC/ AMBT for a 
patient with breast cancer, Pirozzi v. Blue Cross/Blue Shield, 149 was handed 
down in 1990. The plaintiffs plan excluded coverage for treatment that was 
an "experimental or clinical investigative procedure," but neither defmed 
these terms nor reserved judgment about their application for the plan 
administrator.15° Construing what it regarded as ambiguous language against 
the insurer, the court credited the testimony of the treating oncologist that the 
procedure was Pirozzi's best chance for meaningful survival. As in Dozsa, the 

series are subject to notorious selection bias. Medicare Coverage Advisory Comm., supra note 
7. For the Medicare Coverage Advisory Committee's decision memo on HDC/ABMT for 
myeloma, seehttp://www.cms.hhs.gov/ncdr/memo.asp?id=10. For Medicare's ultimate decision 
to approve coverage in 2000 for a subset of myeloma patients, after review of data from one 
randomized trial, see http://www.cms.hhs.gov/mcd/viewncd.asp?ncd_id=35-
30.1&ncd_version=5&show--all (last visited Feb. 8, 2004). 

148. Dozsa, 716 F. Supp. at 138. 
149. Pirozzi v. Blue Cross/Blue Shield ofVa., 741 F. Supp. 586 (E.D. Va. 1990). 
150. /d. at 588. 



246 INDIANA HEALTH LAW REVIEW [Vol. 1:213 

court reasoned that if the care was being used by major centers, it could not 
be investigational. 151 fu the next few years, decisions of this kind proliferated. 
fu 1991, district courts in Connecticut, 152 Maryland, 153 Missouri, 154 and New 
Y ork155 ordered coverage for HDCI AMBT. Over succeeding years, they were 
joined by the Second Circuit, 156 the Ninth Circuit, 157 and other federal courts 
in Alabama/58 the District of Columbia, 159 Tilinois, 160 fudiana, 161 Missouri, 162 

151. /d. at 591. 
152. Bucci v. Blue Cross/Blue Shield of Conn., Inc., 764 F. Supp. 728 (D. Conn. 1991). 

The Bucci court followed Dozsa and Pirozzi in crediting the testimony of treating oncologists 
over the technology assessments of Blue Cross/Blue Shield. It also found "logical, scientifi­
cally," the idea that the therapy would prove successful. Id. at 732. 

153. Adamsv.BlueCross/BlueShieldofMd.,lnc., 757F.Supp.661 (D.Md.1991). Here 
the contract language read ''not generally acknowledged as accepted medical practice by the 
suitable medical specialty practicing in Maryland, as decided by us." ld. at 663. Maryland 
oncologists were typically referring patients to Johns Hopkins for the procedure. 

154. White v. Caterpillar, Inc., 765 F. Supp. 1418 (W.D. Mo. 1991), aff'd, 985 F.2d 564 
(8th Cir. 1991). In White, the court credited the testimony of testifying oncologists over con­
tract language that required following the guidance of reports of the Clinical Efficacy 
Assessment Project of the American College ofPhysicians and the Council on Scientific Affairs 
of the American Medical Association. Neither of these bodies had issued a determination that 
ABMT should be regarded as effective therapy for breast cancer as established. /d. 

155. Kulakowski v. Rochester Hosp. Serv. Corp., 779 F. Supp. 710 (W.D.N.Y. 1991). 
Here, the contract exclusion read: 

We will not provide benefits for any procedure or service which, in the sole 
judgment of the Blue Choice Medical Director, is experimental in nature. In 
addition, we will not provide benefits for medical treatments or procedures not 
proven to be safe and efficacious; nor will we provide benefits for ineffective or 
experimental surgical or medical treatments, or procedures, research studies or 
other experimental health care procedures under continued scientific testing and 
research with questions, in the sole judgment of the Blue Choice Medical 
Director, as to safety and efficacy. 

Id. at 712 n.2. Despite the contract language reserving judgment to the plan, the court con­
cluded that the insurer had applied a new and higher standard, credited the testimony of treating 
oncologists that the therapy was effective, and ruled that the coverage denial had been arbitrary 
and capricious. I d. at 717. The witness the court found most impressive was the oncologist from 
the transplant program to which Kulakowski had been referred; his testimony reported the 
numbers of procedures he had performed and what he regarded as encouraging results from his 
series of patients. !d. at 713-14. The court cited White and Pirozzi in its conclusion. !d. at 715. 

156. Zervos v. Verizon, N.Y., Inc., 277 F.3d 635 (2d Cir. 2002). 
157. Simkins v. Nevadacare, Inc., 229 F.3d 729 (9th Cir. 2000). 
158. Calhoun v. Complete Health, Inc., 860 F. Supp. 1494 (S.D. Ala. 1994). 
159. Wilson v. Group Hospitalization & Med. Servs., Inc., 791 F. Supp. 309 (D.D.C. 

1992). 
160. Frendreis v. Blue Cross Blue Shield of Mich., 873 F. Supp. 1153 (N.D. lll. 

1995); Duckwitz v. Gen. Am. Life Ins. Co., 812 F. Supp. 864, 867 (N.D.lll. 1993) (holding 
HDC/ABMT not an "organ transplant'' within the limits of plaintiff's policy). 

161. Smith v. Office of Civilian Health & Med. Program of the Unif. Servs., 884 F. Supp. 
303 (S.D. Ind. 1994), rev'd, 91 F.3d 905 (7th Cir. 1996). 

162. Henderson v. Bodine Aluminum, Inc., 70 F.3d 958, 960 (8th Cir. 1995) (holding that 
it would violate the Americans with Disabilities Act to cover effective therapy for some cancers 
but not effective HDC/ABMT for plaintiff's breast cancer). 



2004] MANDATED COVERAGE OF EXPERIMENTAL THERAPY 247 

New York, 163 North Carolina, 164 South Carolina, 165 Tennessee, 166 Virginia, 167 

and Washington. 168 Additionally, the Eleventh Circuit ruled that "experimen­
tal" was ambiguous as a matter of law and reversed and remanded a grant of 
summary judgment for the insurer, requiring a trial on the merits. 169 

These decisions typically followed Dozsa in crediting the testimony of 
treating oncologists over insurance companies' contentions that the therapy 
had not been validated scientifically. Courts seemed impressed by the fact 
that the treating oncologists had examined the patients, whereas the experts 
testifying for insurance companies typically had not examined the patients.170 

The courts relied on the observation that major medical centers such as Johns 
Hopkins and Duke were offering the procedure and that community oncolo­
gists were referring their patients to these centers. Some courts explicitly 
borrowed the language typically used to state the standard of care in medical 
malpractice litigation-in the words of one court, whether the care is "gener­
ally acknowledged as accepted medical practice by the suitable medical 
specialty practicing in Maryland."171 Even when the language of the insurance 
contract granted unfettered discretion to plan administrators, some courts 
continued to find the denials arbitrary and capricious, in light of what they 
viewed as the patient's dire situation and the testimony of treating oncologists 
that the experimental treatment was the patient's only hope. Cases in this vein 
were decided as late as 2001.172 Sympathetic to patients, courts sometimes 
went to great lengths to read plan language against insurers. One particularly 

163. Whitneyv. EmpireBlueCross&Blue Shield, 920F. Supp. 477,480(S.D.N.Y. 1996) 
(finding plaintiff benefited substantially from the treatment in February 1993, despite her death 
in July 1993; ordering coverage despite language reserving sole discretion to the insurer and a 
requirement of demonstrated benefit in the peer reviewed literature); Scalamandre v. Oxford 
Health Plans, Inc., 823 F. Supp. 1050, 1057 (E.D.N.Y. 1993) (HDC/ABMT "successfully 
extended [plaintiffs] life expectancy," from July 1991 to her death in January 1992); Kekis v. 
Blue Cross & Blue Shield of Utica-Watertown, Inc., 815 F. Supp. 571 (N.D.N.Y. 1993) 
(following Kulakowski, Dozsa). 

164. Hawkins v.MailHandlersBenefitPlan, No.1:94CV6, 1994 U.S.Dist. LEXIS 20671 
(W.D.N.C. Jun. 28, 1994). 

165. Gripkey v. Mail Handlers Benefit Plan, No. 3:94-378-0, 1994 U.S. Dist. LEXIS 
20619 (D.S.C. Feb. 14, 1994). 

166. Mashburn v. Mail Handlers Benefit Plan, No. 3:94-0549, 1994 U.S. Dist. LEXIS 
19779 (M.D. Tenn. Aug. 4, 1994). 

167. Wheeler v. Dynamic Eng'g, Inc., 850 F. Supp. 459 (E.D. Va. 1994); Bailey v. Blue 
Cross/Blue Shield ofVa., 866 F. Supp. 277 (E.D. Va. 1994). 

168. Bishop v. Office of Civilian Health & Med. Programs of the Unif. Servs., 917 F. 
Supp. 1469 (E.D. Wash. 1996) (ordering coverage by CHAMPUS). 

169. Dahl-Eimers v. Mutual of Omaha Life Ins. Co., 986 F.2d 1379 (11th Cir. 1993). 
170. See Kulakowski v. Rochester Hosp. Serv. Corp., 779 F. Supp. 710, 713 (W.D.N.Y. 

1991) ("Dr. Stankaitis had never seen or examined Mrs. Kulakowski, but based on a review of 
both her records and applicable medical literature, he opined that HDC/ ABMT as proposed for 
her would be 'experimental' and, therefore, excluded from coverage under the contract."). 

171. Adams v. Blue Cross/Blue Shield of Md., 757 F. Supp. 661, 669 (D. Md. 1991). 
172. Smith v. Newport News Shipbuilding Health Plan, Inc., 148 F. Supp. 2d 637 (E.D. 

Va. 2001). 



248 INDIANA HEALTH LAW REVIEW [Vol. 1:213 

unusual line of reasoning was a Virginia court's determination that an 
insurance contract that excluded high dose chemotherapy "with" stem cell 
rescue could be read to cover the chemotherapy separately from the rescue.173 

Another court reasoned that the procedure was not a "transplant" within the 
ordinary understanding of the term.174 

One of the decisions favoring the plaintiff was particularly noteworthy 
for confusing clinical trials with therapy. In Dahl-Eimers v. Mutual of 
Omaha, the Eleventh Circuit remanded a district court's ruling refusing a 
preliminary injunction to the plaintiff pending a determination of the plain­
tiff's likelihood of success on the merits. 175 The court held that the use of 
"experimental" in contract language was ambiguous as a matter oflaw.176 In 
Dahl-Eimers, the plaintiff sought to enter a Phase IIII trial, and the consent 
form specifically stated that the patient might not benefit from participation.177 

The district court judge had concluded that the procedure therefore came 
within the contract exclusion for "experimental" care.178 The Eleventh Cir­
cuit, however, concluded "experimental" was ambiguous and that the plaintiff 
was entitled to a trial on the merits to determine whether participation in the 
trial came within the contract. Notably, the district court judge had vehe­
mently expressed concerns that the lack of federal funding for patients partici­
pating in clinical trials adversely affected patient care and the development of 
medicalknowledge.179 Dahl-Eimers may also be unique for the district court's 
observation of the possible conflict of interest on the part of oncologists 
seeking to enter patients into studies. 180 

Insurers responded to these decisions in part by changing contract 
language in an attempt to avoid rulings based on ambiguity. One strategy was 
to reserve discretion in coverage decisions to the plan administrator. Another 
strategy was to define "experimental" as applying specifically to the context 

173. Bailey v. Blue Cross/Blue Shield ofVa., 866 F. Supp. 277 (E.D. Va. 1994). The two 
therapies must be conjoined because the high dose chemotherapy oblates the patient's 
functioning bone marrow. /d. 

174. Calhoun v. Complete Health, Inc., 860 F. Supp. 1494 (S.D. Ala. 1994). 
175. See Dahl-Eimers v. Mut. of Omaha Ufe Ins. Co., 986 F.2d 1379 (11th Cir. 1993). 
176. /d. at 1383-84. 
177. Dahl-Eimers v. Mut. of Omaha Life Ins. Co., 812 F. Supp. 1193, 1196 (N.D. Fla. 

1992). 
178. /d. at 1197-98. 
179. In the words of the district court judge: 

It is unfortunate that adequate funding for clinical trials of drugs and treatments 
for catastrophic illnesses is not available from the federal government. But such 
funding is quite limited, and is thinly spread among thousands of competing 
programs involving literally hundreds of various illnesses and diseases. Thought­
ful people need to ponder how this problem can be fairly resolved. 

/d. at 1198. 
180. /d. at 1197. 



2004] MANDATED COVERAGE OF EXPERIMENTAL THERAPY 249 

of a clinical trial.181 A third option was to base approval on determinations by 
various evaluating organizations, such as the clinical assessment project of the 
American College of Physicians, the Council of Scientific Affairs of the 
American Medical Association, 182 HCFA, 183 or the NIH.184 Insurance plans 
provided by the federal government for federal employees were particularly 
likely to be successful in referring coverage decisions to the standards of other 
federal agencies, such as Medicare. Federal plans also sometimes followed 
the Medicare standards in specifically excluding breast cancer as an indication 
for which HDC/ ABMT was covered. 185 Where the exclusion was not explicit, 
however, courts ruled in favor of federally-employed plaintiffs as well. 186 

Despite the criticisms in scientific literature of HDC/ABMT, courts 
continued to order coverage for breast cancer patients seeking high dose 
chemotherapy. The Ninth Circuit ruled in 2000 that a plaintiff could not 
reasonably have understood that an autologous bone marrow transplant was 
a "tissue transplant" within the exclusion language of her contract. 187 A 
federal district court in Virginia granted a preliminary injunction for the 
plaintiff mandating the therapy so that she would not lose the window of 
opportunity for the procedure.188 In granting the injunction, the court dis­
counted the need for Phase ill studies to show an advantage for high dose 

181. Lewisv. Trustmarklns.Co., 182F.3d908(4thCir.1999);Wolfv.Prudentiallns. Co. 
of Am., 50 F.3d 793 (10th Cir. 1995); Harris v. Mut. of Omaha Co., No. IP92-1089-C, 1992 
U.S. Dist. LEXIS 21393 (S.D. Ind. Aug. 26, 1992). 

182. See, e.g., Uhrich v. Caterpillar, Inc., No. 93C5271, 1993 U.S. Dist. LEXIS 16420 
(N.D. Ill. Nov. 17, 1993). 

183. See, e.g., Bechtold v. Physicians Health Plan ofN. Ind., Inc., 19 F.3d 322 (7th Cir. 
1994); Fuja v. Benefit Trust Life Ins. Co., 18 F.3d 1405 (7th Cir. 1994). 

184. Lowery v. HealthChicago, Inc., No. 92C7657, 1994 U.S. Dist. LEXIS 6380, at *3 
(N.D. Til. May 13, 1994). 

185. Hawkins v. Mail Handlers Benefit Plan, No. 1 :94CV6, 1994 U.S. Dist. LEXIS 20671 
(W.D.N.C. June28, 1994) (Mail Handlers plan); Jaecks v. Nat'IAss'n of Letter Carriers Health 
Benefit Plan, No. 93C6855, 1993 U.S. Dist. LEXIS 16876 (N.D.lll. Nov. 30, 1993); Regerv. 
Espy, 836 F. Supp. 869 (N.D. Ga. 1993); Goepel v. Mail Handlers Benefit Plan, No. 93-3711, 
1993 U.S. Dist. LEXIS 13346 (D.N.J. Sept. 23, 1993), vacated by 36 F.3d 306 (3d Cir. 1994); 
Dodd v. Blue Cross & Blue Shield Ass'n, 835 F. Supp. 888 (B.D. Va. 1993); Caudill v. Blue 
Cross-Blue Shield, No. 92-94-CIV-7-F, 1992 U.S. Dist. LEXIS 21448 (E.D.N.C. Aug. 24, 
1992). 

186. See Bishop v. Office of Civilian Health & Med. Programs of the Unif. Servs., 917 F. 
Supp. 1469 (E.D. Wash. 1996); Gripkey v. Mail Handlers Benefit Plan, No. 3:94-378-0, 1994 
U.S. Dist. LEXIS 20619(D.S.C. Feb.14, 1994); Wheelerv. Dynamic Eng' g., 850F. Supp. 459 
(E.D. Va. 1994); Mashburn v. Mail Handlers Benefit Plan, No. 3:94-0549, 1994 U.S. Dist. 
LEXIS 19779 (M.D. Tenn. Aug. 4, 1994); Smith v. Office of Civilian Health & Med. Program 
of the Unif. Servs., 884F. Supp. 303 (S.D. Ind. l994);Hawkins, 1994 U.S. Dist. LEXIS 20671 
(CHAMPUS coverage). In the Gripkey decision, the court concluded that the plaintiff had a 
reasonable likelihood of success on the merits because of the confused state of the decisions in 
other cases! 1994 U.S. Dist. LEXIS 20619, at *10. 

187. Simkins v. Nevadacare, Inc., 229 F.3d 729 (9th Cir. 2000). 
188. Smith v. Newport News Shipbuilding Health Plan, Inc., 148 F. Supp. 2d 637 (E.D. 

Va. 2001). 



250 INDIANA HEALTH LAW REVIEW [Vol. 1:213 

treatment over conventional chemotherapy; it found the testimony of treating 
oncologists sufficient. 189 In 2002, the Second Circuit ordered coverage of the 
therapy for a male patient with breast cancer, despite the patient's health 
plan's (Empire Health Choice) reliance on recent technology assessments.190 

Also in 2002, a federal district court in Michigan rejected an insurer's 
decision that high dose chemotherapy was not indicated for someone with the 
patient's stage and type of disease, although it was covered for other indica­
tions.191 Finally, a state court in Arkansas rejected an insurer's insistence on 
data showing that high dose chemotherapy was superior to standard therapy 
before extending coverage.192 No reported cases in 2003 appear to have ruled 
on the issue. 

To be sure, courts have not ruled uniformly in favor of plaintiffs seeking 
coverage for high dose chemotherapy. Insurance companies have prevailed 
in some cases, especially when contract language was absolutely explicit 
about excluding the procedure for breast cancer. Nonetheless, courts have 
been persuaded by the sympathetic situations of patients believing that court 
intervention was the patient's only hope. They have relied on the testimony 
of referring oncologists rather than on the published scientific literature. In 
so ruling, courts have reinforced the beliefs of patients that the care is both 
beneficial and covered. Courts have thus encouraged patients to form ex­
pectations in reliance on experts without sufficient scrutiny to determine 
whether this reliance is reasonably placed. Like legislatures, courts have not 
sufficiently distinguished therapeutic reasons from other moral reasons for 
providing care to patients when its efficacy is under investigation. 

VI. CONCLUSION 

In mandating coverage for plaintiffs, courts sometimes noted public 
policy concerns. Judges referred to the financing of health care in America 
as being in a state of crisis193 and to the difficult judgments involved in fund­
ing expensive new modalities of care.194 These are indeed difficult questions. 
It is important to try to determine whether new therapeutic modalities carry 
promise. Funding patient participation in clinical trials is a major step in this 
direction. It is also important to ensure that new medical knowledge is 
developed in a manner that is fair to all patients, so that some groups (women, 
for example) are not left out. Funding participation in clinical trials is like-

189. /d. at 652. 
190. Zervos v. Verizon, N.Y., Inc., 277 F.3d 635 (2d Cir. 2002). 
191. Reed v. Wal-Mart Stores, Inc., 197 F. Supp. 2d 883 (B.D. Mich. 2002). 
192. Am. Inv. Life Ins. Co. v. Butler, 65 S.W.3d 472, 474 (Ark. Ct. App. 2002). 
193. Goepel v. Mail Handlers Benefit Plan, No. 93-3711, 1993 U.S. Dist. LEXIS 13346 

(D.N.J. Sept. 23, 1993). 
194. Jaecks v. Nat' I Ass'n of Letter Carriers Health Benefit Plan, No. 93C6855, 1993 U.S. 

Dist. LEXIS 16876 (N.D. Ill. Nov. 30, 1993). 



2004] MANDATED COVERAGE OF EXPERIMENTAL THERAPY 251 

wise important to this end. If it is beneficial for patients to participate in 
clinical trials, justice requires that access to the trials is fair. None of these 
reasons, however, support encouraging patients to believe they are entitled to 
experimental care as though it were already established therapy. Unfor­
tunately, this is what the courts seem to have done in ruling in favor of 
patients seeking high dose chemotherapy for breast cancer. This is also what 
statutes mandating coverage for off-label drug use and participation in clinical 
trials may do if the mandates are not carefully limited. In the case of off.label 
drug use, mandates should be limited to uses that have been scientifically 
validated; some, but not all, of the statutory mandates incorporate this limit. 
In the case of participation in clinical trials, coverage should be limited to 
trials that are likely to yield scientifically helpful results. Otherwise, the result 
may be further resort to courts by patients who have come to expect care as 
beneficial and believe, therefore, that the care is an entitlement. Even if these 
expectations have been encouraged by the law, however, they are not 
legitimate and thus are not grounds for moral entitlement, because the beliefs 
on which they are grounded are not reasonable.