The Cancer Journal - Volume 8, Number 5 (September-October 1995)
Medical ethics is in a dilemma - To support human experimentation, or to ban it? The wealth of publications on this issue testifies to the bewilderment. Medical ethics are "the values and guidelines that should govern decisions in medicine" (1). This definition is too broad, impersonal and therefore disappointing. What are these "decisions in medicine"? Health care, or how to build hospitals? Why not narrow the issue, and focus on the conduct of the physician in his relationship with his patient? This might rescue medical ethics from its bewilderment, since the guidelines are simple and straightforward: "I will prescribe regimen for the good of my patients according to my ability and my judgment and never do harm to anyone" "In every house where I come I will enter only for the good of my patients" (Hippocratic Oath) (1). The ethical dilemma is connected with the randomized clinical trial, a research tool for investigating safety and efficacy of new drugs, particularly in cancer (2). Drugs are evaluated in three phases. The Phase I trial is more concerned with drug toxicity and less with its efficacy. Its main objective is to determine the maximally tolerated dose of a drug, evaluate its toxicity, pharmacokinetics, and to recommend doses for further testing. Unless found toxic, a drug is also evaluated in phase II trial when it is tested for efficacy. Effective drugs are then tested in phase III trials.
Shortcomings of Phase I and II trials - By their very nature clinical trials are concerned with the average patient, otherwise statistical methods, e.g., hypothesis testing, are useless. Yet rarely does an average patient represent the response of the individual, since variability of patient responses to treatment is considerable (3). Even if treatment is adequate for the average patient other patients may not benefit from it, and very few patients achieve partial or complete response during phase I trials (4). The statistician Austin Bradford Hill summarized this difficulty concisely: "The clinical trial does not tell the doctor what he wants to know" (5). Most patients participating in phase II trials do not respond to the recommended dose that was established in the phase I trial. Overall response rate for new agents in phase II trials was 4% which is less than the overall objective response rate in phase I trials of 6% (6). 96% patients did not benefit at all.
Shortcomings of Phase III trials - A properly designed phase III trial should prove or disprove the null hypothesis with high probability. Which is hopefully achieved by the following means: 1: Increasing the number of participants; 2: Improving group homogeneity, by matching different attributes, e.g., age, gender, race, performance status, socio-economic status, tumor grade and stage; 3: Selecting a well defined trial end point, e.g., survival, or relapse-free interval. These requirements are hardly ever met, particularly group homogeneity. Before assigning patients to different treatment arms their attributes, e.g., age, socio-economic status have to be matched, which is not at all simple. A correct matching of their birth dates does not match their biological age, which determines their response to therapy. Or, how does one match the socio-economic status of two individuals? Since tumor stage is poorly correlated with prognosis, stage matching does not assure that the conditions of patients in the two treatment arms are the same. Such inadequacies may arbitrarily enrich one treatment arm with sicker patients and undermine the reliability of the trial.
Ethics of clinical trials - Now suppose that you are asked to advise your patients to participate in Phase I and II trials. Would you advise them to participate, knowing that 96% will not benefit or may even be harmed by the treatment? (6). Or would you advise your patient to participate in a Phase III trial that compares the effectiveness of a new and unknown drug that might be worse than the drug that your patient requires? Your oath forbids you to do so! Yet modern medicine advocates clinical trials, facing a profound ethical dilemma, when is human experimentation allowed? The Declaration of Helsinki is an attempt to define such guidelines (7): "In any medical study, every patient - including those of a control group, if any - should be assured of the best proven diagnostic and therapeutic method". This sounds like a modern version of the Hippocratic oath. Why then are clinical trials allowed? The quest for their justification led to the proliferation of publications on medical ethics. Some are pure hypocrisy, stating that: "A well designed clinical trial offers the patient access either to the best standard therapy or to a new treatment that is, at the outset, considered at least equivalent or possibly superior to the standard treatment" (8). Others narrow the scope of the problem, suggesting that only placebos in clinical trials are unethical, ignoring Phase I and II trials that do not use placebos (7).
Informed consent - Efforts to justify clinical trials culminated in informed consent. The patient has to agree to participate in the clinical trial and understand what he is up to. Which patient would agree to participate in a trial in which his chances of benefit are 4%? Yet they agree, since they trust their physician. By signing the informed consent form they relieve the physician from the ethical burden. It is an "escape clause that puts the burden on the patient for ethical studies" (9). Some believe that the Declaration of Helsinki is flawed since it does not allow the patient to accept small risks of discomfort (9). Why not ask terminally ill patients to serve as experimental objects for the benefit of society and future generations? After all, the state that sends its sons to die in the battlefield, should also be allowed also to force its other sons to be experimented upon...If this sounds like a recent nightmare, you will appreciate the Declaration of Helsinki document.
Ethics are relative - Medicine has adopted two ethical codes: traditional ethics that direct the conduct of the physician with his patient, and academic ethics, that promote human experimentation. Both exist side by side, which may disturb philosophers who still believe in absolute ethics. One ethical discipline is concerned with the individual, while the other with the collective. Academic medicine has adopted "collective ethics" since its concept of the normal is based on the collective average (10, 11). The mean blood sugar of the healthy population is regarded as normal blood sugar. Any deviation is pathological, e.g., hypo- or hyper-glycemia. New drugs are therefore tested on the collective, and these needs created a the new ethical discipline that is called here academic ethics. In his important work "The Normal and the Pathological" (10, 11), Georges Canguilhem, distinguishes between two types of normal: individual normal, or norm, and collective normal, the statistical average. Only the first is meaningful, and since Canguilhem rejects the collective normal, he should not be bothered by academic ethics. Collective normal is meaningless since it cannot be applied to the individual. This is why Austin Bradford Hill concluded that the clinical trial "does not tell the doctor what he wants to know" (5).
Human experimentation - Nevertheless, human experimentation is essential for good medical treatment. Every drug has its side-effects, and since treatment outcome is unknown, it is by nature experimental. If the physician encounters a new drug that seems to helpful to his patient, he ought to test it, because he has sworn that, "I will prescribe regimen for the good of my patients according to my ability and my judgment and never do harm to anyone" (1). Still, judgment may vary, and even the best physician may be wrong. How to protect the patient from malpractice? This issue can be solved outside the scope of medical ethics, e.g., by quality control.
1. Dorland's Medical Dictionary. 28th Edition. WB Saunders Co. Philadelphia, 1994.
2. Kaufman D. Cancer therapy and the randomized clinical trial. Good medicine? Cancer Suppl. 72, 2801-2804, 1993.
3. Ratain MJ, Mick R., Schilsky RL, Siegler M. Statistical and ethical issues in the design and control of phase I and II clinical trials of new anticancer agents. J. Nat Cancer, Inst. 85 , 1637-1643, 1993.
4. Estey E, Hoth D, Simon R. et al. Therapeutic response in phase I trials of antineoplastic agents. Cancer Treat. Rep. 70 ,1105-1115,1986.
5. Horwitz RI. Large-scale randomized evidence: large simple trials and overviews of trials: a clinician's perspective on meta-analyses. J. Clin. Epidemiol 48, 41-44,1995.
6. Von Hoff DD, Turner J. Response rates, duration of response and dose response effects in phase I studies of antineoplastics. Invest New Drugs 9 , 115-122, 1991.
7. Rothman KJ, Michels KB. The continuing unethical use of placebo controls. New Engl. J. Med. 331394-398, 1994.
8. Levine RJ. Ethics of clinical trials. Do they help the patient? Cancer Suppl. 72, 2805-2810, 1993.
9. Taubes G. Use of placebo controls in clinical trials disputed. Science 267 ,25-26 ,1995.
10. Canguilhem G. Le Normal et le Pathologique. translated into English by Fawcett CR, Cohen RS. Zone Books New York 1991.
11. Zajicek G. The normal and the pathological. Cancer J. 7, 48-49, 1994.