Who Should Go First in Trials with Scarce Agents? The Views of Potential Participants

Demands from AIDS activists in the 1980s for access to drugs not approved by the U.S. Food and Drug Administration (FDA) resulted in multiple programs to increase access, including single-patient treatment Investigational New Drug (IND) applications, treatment INDs for groups of patients, and a streamlined approval process for promising agents.1 Cancer patients have benefited from these expedited processes. For example, imatinib mesylate was approved for chronic myelogenous leukemia in less than three months using the FDA Fast Track program and was given to more than 5,000 nonclinical trial patients in an expanded access program (EAP) before final FDA approval. Similarly, gefitinib was approved in less than 10 months for non-small-cell lung cancer and was made available to more than 37,000 patients in an EAP.2 Yet calls for even greater access continue from medical researchers, clinicians, and patient advocacy groups,3 including a legal action alleging that terminally-ill patients have a liberty right to access experimental treatments.4
Competing rights—specifically “the individual right to self-determination and the right of the state to effect some form of control to protect the public and advance science”5 —complicate the issue of access to nonvalidated agents. Advancing science is predicated upon the promotion of aggregate rather than individual health6 and seemingly subordinates the interests of individuals, some of whom may be terminally ill. Referring to patients who died pursuing access to investigational agents, Burroughs argues that patients without therapeutic options lack the time to wait for science to discover, test, and validate new treatments.7 Patients may consider that any potential therapy—including an investigational agent with unknown risks and efficacy—is worth a try when weighed against imminent death. Proponents of increased access to investigational drugs argue that barriers interfere with an individual’s right to determine for oneself the acceptable balance between the burdens and benefits of participating in research.

Opponents warn that open access poses risks to dying patients, such as diminished quality or even length of life,8 as well as promoting a therapeutic misconception—the mistaken belief that research protocols are intended primarily to benefit individual subjects. Moreover, expanding access risks diverting resources from clinical trials to the treatment of patients, thereby losing valuable information and time in the search for safe and effective new therapies9 and undermining the potential benefits of research for society.

We assume that patients’ views and experiences should inform the debate about the appropriate way to provide access to investigational agents. Accordingly, we reviewed survey responses from patients who were being considered for a phase I clinical trial of a widely heralded antiangiogenesis agent about their views on access to investigational agents.

Study Methods

We used information from a survey conducted with 100 of the first 130 persons with solid tumors who had been referred to a phase I trial of human recombinant endostatin at The University of Texas M.D. Anderson Cancer Center between October 1999 and October 2000. The 30 nonresponders were characterized previously.10 The survey instrument was based on the Advisory Committee on Human Radiation Experiments’ Subject Interview Study Brief Survey in consultation with one of its designers (JS). In addition to nine closed-choice and one open-ended question about access to investigational agents listed in Appendix I, the survey also inquired into the influence of media coverage on patients’ understanding of the purpose of this phase I trial,11 attitudes toward research, quality of life,12 and demographic characteristics. Oral consent was obtained for surveys conducted in person or by phone, and an informed consent statement prefaced self-administered surveys. The endostatin trial and the survey were approved by the Institutional Review Board at M.D. Anderson.
Descriptive statistics were used to summarize the responses to the closed-choice questions. For each closed-choice question, Chi-square test in Minitab was used to test for significance. To better understand the overall results, we aggregated results for consistent responses to three items that asked whether access to investigational drugs was too easy, too hard, or just right into a single measure. The one participant who did not give a consistent view, answering yes to both “Too Hard” and “Just Right,” was labeled as “Do Not Know” in the combined response. The open-ended question, “If [investigational] drugs are only to be given to a few people, what is the fairest way of deciding who gets them?”, was examined using content analysis. One author (RP) identified labels to match the content by examining responses line by line. A second author (AF) read all the participants’ answers and refined the content labels. AF and RP used the final category list to independently classify each respondent’s answer, and MC resolved any disputes. Fisher’s Exact test in SAS was used to analyze whether there are significant relationships between allocation criterion and the following covariates: whether respondents understood the purpose of the trial, whether the survey was conducted before/after the informed consent process, and whether respondents were aware of their trial status (aware on/off trial or not aware—that is, waiting to find out if they would be offered a place on the trial). We counted three types of responses as evidencing a respondent’s correct understanding of the purpose of a phase I trial: 1) monitoring drug safety, toxicity, and/or side effects; 2) dose determination; and 3) learning whether the drug works or cures cancer, if the respondent also mentioned safety or dose determination. Although the primary purposes of a phase I trial are safety and dose determination, efficacy is monitored. Therefore, we considered the mention of efficacy correct as long as the participant also mentioned safety or dose determination. Respondents’ preferences for the two most commonly cited allocation criteria (offering investigational drugs to those who would benefit most or were most needy versus allocating to maximize scientific advancement) were tested given the above-noted three covariates, as well as respondents’ age, education (some college versus no college), income (less than $60,000/year versus greater than $60,000/year), gender, and ethnicity. A multivariable logistic regression analysis was conducted in SAS to find which covariates could be a predictor of the need/benefit or science allocation criterion.


The median age of the 100 respondents was 56 years (range, 25-79 years). Eighty-six percent were white, 7% Latino, 4% African-American, and 3% other or missing. Sixty-nine percent of the respondents had some college education; 17% of these had graduate degrees. Over half of the respondents reported income of more than $60,000 a year. Ninety-three percent had some form of medical insurance (81% private and 35% public; some reported having both types of insurance). Most respondents (77%) were not motivated by a belief that participation was the best way to pay for cancer treatment.
Respondents were more likely to agree than disagree that patients had a right to investigational drugs (p < 0.001), that investigational drugs should be offered only in the context of research (p = 0.006), that access to investigational drugs was too hard (p < 0.001), and that knowing the right people (p = 0.014) and being persistent (p < 0.001) increased one’s chance of access. Respondents did not agree that investigational drugs should be given to anyone who wanted them (p = 0.006) or that physicians were aware of the latest ones (p < 0.001) (Table 1). The overall combined responses to the ease of access questions demonstrated that responders were more likely to agree that access to investigational drugs was too hard than to agree that access to investigational drugs was just right (p < 0.001) (Table 2).

In response to the open-ended query about how to fairly distribute scarce investigational drugs, respondents most frequently identified two allocation criteria: 1) offering investigational drugs to those who would benefit most or were most needy; and 2) allocating to maximize scientific advancement. Respondents who understood the purpose of trial preferred the second criterion (p = 0.002). No significant relationship was found between the allocation criteria and whether the survey was conducted before or after the informed consent process was initiated, nor whether respondents were aware of their trial status (Table 3). We found no trend distinguishing the views of the 80 respondents who did not know their trial status from the views of the 20 who knew their status.

In a multivariable logistic regression analysis, we did not detect significant associations between the other characteristics of the respondents (age, race, income, education, or medical insurance) and choice of allocation criterion. Only correct understanding of the purpose of the trial was found to be significantly related to the choice to allocate to maximize scientific advancement. (estimated odds ratio = 3.33, 95% C.I. = [1.07, 10.37], p = 0.035) (Table 4).

Responses endorsing need as an allocation criterion reflected several nuances, including prioritizing those with no therapeutic options (“If [the person has] been through therapy and it’s unsuccessful, it’s more important to get that person on investigational therapy”) and those who had more advanced disease or were close to death (investigational agents should go to “those closer to dying; those who really need something now”). Several respondents parsed need in terms of responsibilities: “I have two children. I’m 53. My husband is already dead. So I am in the middle here. I’d sacrifice my life to save a 32-year-old dad with a four-year-old.”

Responses favoring scientific criteria were often explicit: “Criteria should be what will best establish the efficacy of the drug. If [subjects are] assigned according to patient need, and whether the patient stands to benefit, results will be hazy. Scientists won’t know whether to make [a] drug available or not.” Others explained that science-driven criteria should include the ability to complete the trial. As one respondent commented, trial participants should be “healthy and hearty enough with not many other complications, so [investigators] can tell if the drug is effective. [Participants should have] availability, willingness to participate, willingness to go through the routine, mental state/positive attitude.” Eight responses combined both need and science criteria, for example: “Give it to the people who would most benefit and where the investigators will get the most information about the drug.”

In addition to their primary answer, four respondents mentioned that younger people should receive priority. One participant combined need with younger age: “The people who have no standard therapy left [should be offered investigational drugs]. Aggressive cancer patients first—slow growing cancers a lower priority. Children and teenagers [should be given] more opportunity over [people in their] 60s and 70s.”

Twelve respondents criticized particular allocation methods. Most of these warned against enrolling patients who were so sick they could not complete the trial and would thus thwart the scientific endeavor. Several criticized random selections: “Lottery is insane; seems inefficient. Have to have a scientific approach, some criteria” and “Not sure lottery is fairest, though it is least subjective. Perhaps should be based more on possible outcome of taking [the drug], or immediate need for patient.” Some respondents suggested that a random procedure could be used as a secondary allocation determinant if there were too many potential participants who met the science criterion (four respondents) or the need criterion (one respondent).

Finally, several respondents refused to suggest an allocation scheme. As one responded, “I don’t know, because I believe that anyone willing to take it should be entitled to take it.” Another said, “Companies that make drugs should make enough—make more. I don’t care if it’s FDA approved.”


Respondents’ views about access to investigational agents reflect the core ethical tension between maximizing scientific advancement and making investigational agents available to ailing individuals. These respondents overwhelmingly stated that hope for personal benefit drove their willingness to participate in a phase I trial,13 yet a high percentage favored a science-based allocation criterion and access restricted to participation in clinical trials, revealing a coincident acknowledgment of the importance of scientific goals. Correct understanding of the purpose of a phase I trial had a significant effect in determining their choice of the scientific-based allocation criterion.
Of course, our results should be interpreted with some limitations in mind. As a single-institution survey at a research facility, our sample, with its high socioeconomic characteristics, may be biased in favor of access to research, and the results may not be generalizable to all potential research participants. In addition, the endostatin trials received an unusual amount of media attention, and this attention may have increased the respondents’ appreciation for the scientific endeavor. (As we reported earlier, those who first heard about endostatin from the media were more likely to understand the purpose of the trial.)14 Further, even though our analysis did not find any differences among the 80 who did not know their trial status and the 20 who did, it still might be possible that the 80 tailored their answers in ways they thought would increase their chances of being admitted to the trial, even though the survey team assured all respondents that their answers would not be communicated to the researchers and that the survey team had no input in the selection process. The survey team identifying themselves as ethicists may have colored respondents’ answers as well. It is also possible that respondents were not aware of allocation alternatives such as randomization or first come, first served and would have preferred them had they been aware. Only a few respondents (13.2%) mentioned such alternatives in response to the open-ended question.

Regardless of these limitations, much can be learned from the data reported here. Over half of the respondents preferred that need or chance of benefit should be the primary or governing factor in offering access to investigational drugs. This is consistent with several existing allocation schemes that weigh patients’ needs along with safety and scientific validity. The FDA Modernization Act of 1997 stipulates that Individual Patient INDs are to be issued for “a serious disease or condition” for which there is “no comparable or satisfactory alternative therapy.”15 Treatment INDs, which target groups of patients, have a similar criterion, but add that “in the case of serious diseases, there [should be] sufficient evidence of safety and effectiveness to support such use” and for “immediately life threatening diseases, available evidence [should] suggest the drug or device may be effective for its intended use and would not expose patients to unreasonable risk” (emphasis added).16 According to these FDA standards, the more life-threatening the disease (resulting in a higher level of need), the lower the standard of efficacy required. This same principle of calibrating risk to need is evident in the FDA’s latest proposed rule change intended to increase access to investigational agents.17 After analyzing how access to gene transfer trials is determined, Modell concluded, “Even if one believes that choosing those subjects best for research is also best for society, practical events as they have unfolded demonstrate that the societal preference lies in mixing scientific progress criteria with other factors like urgency.”18

Using urgency of need as an allocation criterion does not necessitate holding a therapeutic misconception. One may accept that a drug in the early stages of research lacks proven efficacy and still believe that those who have no therapeutic alternatives should be offered first access from a pool of eligible participants. Just as trial participants’ realism can coexist with hope,19 the search for personal benefit can coexist with sound scientific design and benefit to society. However, to decrease the chance of therapeutic misconception and to shield participants from disproportionate harm, an allocation plan based on need or benefit should consider requiring an objective and independent healthcare advisor to evaluate and discuss with prospective participants all available options before an offer of trial participation is made. Group Health Cooperative of Puget Sound used such a procedure in the 1990s to assist, counsel, advise, and, on occasion, dissuade patients who demanded bone marrow transplantation for breast cancer, an often-cited example of the hazards associated with concession to demands for investigational drugs and investigational procedures before efficacy is established (personal communication from Aubrey Davis, CEO Emeritus, Group Health Cooperative). Advisor healthcare providers are likely to be better positioned to ask critical questions about efficacy and side effects that the patient may not know enough to ask.

When asked to describe the fairest allocation method, these respondents did not spontaneously suggest objective distribution systems incorporated in some current approaches. Ethicists have endorsed systems such as “first come, first served” or lotteries.20 Several programs have relied on random systems for distributing scarce investigational drugs. Imclone’s EAP randomly distributed cetuximab to 30 patients a month in 2003. The Children’s Oncology Group Phase I Consortium formerly used a random procedure to assign priority numbers to eligible participants on a waiting list. Another endostatin phase I trial center used a random process to select participants, stating “we have done everything possible to make our procedures open, impartial and objective.”21 If further research demonstrates that participants prefer need- or benefit-based allocation schemes to objective ones, this preference could inform future policy so that need- or benefit-based allocation systems replace or supplement objective ones.

Several respondents mentioned participants’ ability to withstand the demands of a trial as an optimal access criterion, but no respondent mentioned the need for a criterion that prevented disproportionate harm. Perhaps they assumed that investigators, sponsors, and oversight agencies would ensure safety, but the omission hints that, as Agrawal and Emanuel22 have hypothesized, phase I candidates may not focus on the risks posed by new agents—at least not in the same way as do (healthy) observers of the early trial process. Yet preventing undue harm remains the countervailing reason regulatory agencies preclude broader access. EAPs typically do not release drugs until after phase I toxicity testing. As we already noted, the FDA requires some evidence that the risk is not unreasonable before approving an EAP. We agree with this emphasis on patient safety; despite some patients’ willingness to “die fighting,” the ethical principles of beneficence and nonmaleficence support excluding patients who face disproportionate harm from trial participation.23 For example, the M.D. Anderson endostatin trial excluded patients if the tumor biopsy posed significant risk to the patient, even if the patient wanted to accept that risk.

In sum, our respondents’ views are consistent with the current trend to make investigational agents more widely available. However, public policy must balance responsiveness to the public, patient safety, and the integrity of the drug development process.24 Once participant preferences are established by more extensive research, we recommend that the future design of allocation schemes reflect them, but only as long as safety is not compromised and scientific validity is retained.


This research was supported in part by a National Cancer Institute grant U01 CA62461.

An early analysis of these data was presented at the 2002 Office of Research Integrity Conference on Research Integrity, Potomac, Maryland: Pentz RD, Flamm AL, Sugarman J, Cohen MZ, Ayers GD, Herbst RS, Abbruzzese JL. Research access: The views of prospective research participants.

Rebecca D. Pentz, PhD, is Professor of Hematology and Oncology in Research Ethics, Winship Cancer Institute, Emory University, Atlanta, GA;Anne L. Flamm, JD, is Visiting Assistant Professor, Case University School of Law, Cleveland, OH;Jeremy Sugarman, MD, MPH, MA, is the Harvey M. Meyerhoff Professor of Bioethics and Medicine at Berman Institute of Bioethics and Department of Medicine, The Johns Hopkins University, Baltimore, MD;Marlene Z. Cohen, PhD, RN, FAAN, is John S. Dunn, Sr., Distinguished Professor in Oncology Nursing and Assistant Dean and Chair, Department of Integrative Nursing Care, The University of Texas Health Science Center Houston School of Nursing, and Professor, Department of Symptom Research, University of Texas M.D. Anderson Cancer Center, Houston, TX;Zhiheng Xu, MS, is Biostatician, Department of Biostatistics, Rollins School of Public Health, Emory University, Atlanta, GA;Roy S. Herbst, MD, PhD is Chief, Section of Thoracic Medical Oncology, and Codirector, Phase I Clinical Trials Working Group, University of Texas M.D. Anderson Cancer Center, Houston, TX; andJames L. Abbruzzese, MD, is Chair, Department of GI Medical Oncology, University of Texas M.D. Anderson Cancer Center, Houston, TX.


1. Levine R. The impact of HIV infection on society’s perception of clinical trials.Kennedy Institute of Ethics Journal1994;4:93-98; Weijer C. Evolving ethical issues in selection of subjects for clinical research.CambridgeQuarterly of Healthcare Ethics1996;5:334-345.

2. Reiling RB, Natale R, Wade J, et al. Efficacy and safety of gefitinib in chemo-naive patients with non-small cell lung cancer (NSCLC) in an expanded access program. In 2005 ASCO Annual Meeting Proceedings.Journal of Clinical Oncology2005;23(16S Pt I):7094.

3. Medicare coverage of clinical trials: Testimony of Jan Platner, JD, Director of Administration and Programs, National Breast Cancer Coalition before the Agency for Healthcare Research and Quality (November 20, 2000),http://www.natlbcc.org/bin/index.asp?strid=378&depid=3; National Cancer Institute. Cancer clinical trials: A new national system. Broadening access for physicians and patients. (November 17, 2003),http://www.cancer.gov/clinicaltrials/conducting/new-national-system/page4; Availability of investigational drugs for compassionate use. Statement by Robert J. Temple, MD, Associate Director for Center for Drug Evaluation and Research, Food and Drug Administration before the Committee on Government Reform, 106th House of Representatives. (June 20,2001),http://www.fda.gov/ola/2001/compassionateuse0620.html.

4.Abigail Alliance for Better Access to Developmental Drugs and Washington Legal Foundation v. Von Eschenbach and Levitt, 445 F.3d, 470 (US App. 2006).

5. Salisbury DA, Schechter MT. AIDS trials, civil liberties and the social control of therapy: Should we embrace new drugs with open arms?Canadian Medical Association Journal1990;142(10):1061.

6. Wendler D. When should “riskier” subjects be excluded from research participation?Kennedy Institute of Ethics Journal1998; 8:307-327.

7. Burroughs F, Walker S. Response to submission from Ms. Musa Mayer et al. regarding “Tier 1 Initial Approval.” Letter (October 7, 2003),http://abigail-alliance.org/AbigailAllianceReplytoMusaMayerletter-Oct7.pdf.

8. Visco F. NBCC sends letter to FDA—response to Abigail Alliance’s Citizens Petition. Letter (September 9, 2003),http://www.stopbreastcancer.org/bin/index.asp?strid=622&btnid=5&depid=3.

9. Richman D. Public access to experimental drug therapy: AIDS raises yet another conflict between freedom of the individual and welfare of the individual and public.Journal of Infectious Disease1989;159:412-415.

10. Pentz RD, Flamm AL, Sugarman J, et al. Potential participants’ perceptions of a high-profile phase I trial.Journal of Clinical Oncology2002;20:3785-3791.

11. See ref. 10, Pentz et al. 2002.

12. Cohen MZ, Slomka J, Pentz RD, et al. Phase I participants’ views of quality of life and trial participation burdens.Supportive Care in Cancer2007: published online January 25, 2007.http://www.springerlink.com/content/c209651486930352/?p=fa4d0b85a46e4be6b43c.

13. See ref. 10, Pentz et al. 2002.

14. See ref. 10, Pentz et al. 2002.

15. Food and Drug Administration Modernization Act of 1997, Section 561(b) 1.http://www.fda.gov/CDER/guidance/105-115.htm.

16. Food and Drug Administration Modernization Act of 1997, Section 561(c)6-7.http://www.fda.gov/CDER/guidance/105-115.htm.

17. Food and Drug Administration. Expanded access to investigational drugs for treatment use: Proposed rules.Federal Register2006;71(240):75147.

18. Modell S. Framework for the selection of subjects in human gene therapy and diagnostic trials.Journal of Investigative Medicine1996;44:235.

19. Agrawal M, Emanuel E. Ethics of phase I oncology studies: Reexamining the arguments and data.JAMA2003;290:1075-1082.

20. Levine C, Dubler N, Levin RJ. Building a new consensus: Ethical principles and policies for clinical research on HIV/AIDS.IRB: A Review of Human Subjects Research1991;13:1-17; Macklin R, Friedland G. AIDS research: The ethics of clinical trials.Law, Medicine and Health Care1986;14(5-6):273-280.

21. Enrollment process for first human study of Endostatin protein begins Tuesday. Press release (September 28, 1999).http://www.danafarber.org/abo/news/pressarchive/092899.asp.

22. See ref. 19, Agrawal et al. 2003.

23. See ref. 6, Wendler et al. 1998.

24. See ref. 17, Food and Drug Administration, 2006.

Rebecca D. Pentz, Anne L. Flamm, Jeremy Sugarman, Marlene Z. Cohen, Zhiheng Xu, Roy S. Herbst, and James L. Abbruzzese, “Who Should Go First in Trials with Scarce Agents? The Views of Potential Participants,”IRB: Ethics & Human Research29, no. 4 (2007): 1-6.