Intended for healthcare professionals

Education And Debate

How can research ethics committees protect patients better?

BMJ 2003; 326 doi: (Published 29 May 2003) Cite this as: BMJ 2003;326:1199
  1. Silvio Garattini, director (garattini{at},
  2. Vittorio Bertele, head of laboratory of regulatory policies1,
  3. Luca Li Bassi, consultant, laboratory of regulatory policies1
  1. 1 Mario Negri Institute, Via Eritrea 62, 20157 Milan, Italy
  1. Correspondence to: S Garattini
  • Accepted 15 May 2003

The duties of research ethics committees are becoming increasingly difficult—what skills and knowledge do their members need to evaluate protocols that contain elements that are not in the patient's interests?

Drug development and marketing has become a huge industry. The potentially enormous profits mean that to comply with their duty and act in the interest of patients and public health, research ethics committees have to carefully assess clinical protocols with the necessary scientific knowledge and evaluate the true objectives of new studies and the methods used.

Embedded Image


We have selected four main problems that such committees should examine critically.

The use of placebos

The latest revision of the Declaration of Helsinki has reinforced doubts about whether placebos are always used correctly.1 Using a placebo instead of a comparator drug is advantageous for the sponsor because the new drug has more chance of being superior, and this superiority can be shown with fewer patients and therefore with less expense. The use of a placebo can never be justified when a drug of proved efficacy is already available for a given therapeutic indication.

Some justify the use of placebo by stating that it can be used in addition to a comparator (a three arm trial) because the difference between the placebo and the comparator forms the basis for validating the result with the new drug.2 3 This can be challenged on several grounds. Firstly, it is not ethical to deprive patients of useful treatment even if the trial lasts only a few weeks. Secondly, three arm trials are not essential to validate results, as using a comparative trial of sufficient size would make false negative results unlikely. Thirdly, if the aim is to establish the usefulness of the new drug for patients, it is logical to test it on patients who are resistant to the existing treatment. It would then be perfectly acceptable to randomise patients to placebo or the new drug, which would thus have a harder test than the first members of the therapeutic group.

Clinical trials on resistant patients are rare. The notable exception is anticancer agents, which are, however, often tested without controls and randomisation. Despite the alleged ethical reasons, as these drugs are used in patients with advanced cancer the aim is usually to avoid placebo and to obtain a restricted indication easily, which can then be enlarged by promoting off-label use.

The use of placebo must therefore be soundly justified to exclude the possibility of alternative experimental schemes.

Equivalence and non-inferiority trials

In line with the general commitment to ask important questions and answer them reliably4 a trial should be designed to show the superiority of a new treatment over the best treatment available for the same therapeutic indication, the “gold standard.” Superiority should not necessarily be measured only in terms of efficacy, adverse reactions or altered compliance due to easier administration should also be considered.

Summary points

Drug development and marketing has become a multibillion dollar industry, and financial interests are bound to influence the design and planning of clinical trials

Research ethics committees are an essential component of a system that should protect individual patients as well as public health by critically appraising the scientific and ethical content of clinical trials

Committee members may not always possess the knowledge and skills necessary to fulfil these functions, and their duties are becoming increasingly difficult, especially because of the complex design of multicentre international trials

Ethics committees should concentrate on key aspects that make a new clinical trial ethically and scientifically justifiable: the use of placebo; comparator and doses of drugs to be tested; equivalence and non-inferiority trials; and the choice of appropriate end points

As most trials are commercially sponsored, ethics committees should help to minimise the non-publication of negative results—for example, they may say it is unacceptable for a trial protocol to allow the sponsor to veto publication over researchers who bear the scientific responsibility

In recent years superiority trials have often been replaced by equivalence or non-inferiority trials. This reflects a switch from the search for better drugs to the acceptance of drugs that are similar to, or not worse than, those already on the market. This shift is unethical for several reasons.

Firstly, it is conceptually difficult to establish the limits (excess of outcome events) that define a drug as equivalent or not inferior. For drugs with important effects—for instance, life saving—what can we reasonably accept? Is a 10%, 5%, or even 2% difference acceptable as equivalent in the interest of patients? The case has been made for fibrinolytic agents,5 but this question is also relevant for several classes of drugs that reduce morbidity or mortality, such as statins, angiotensin converting enzyme (ACE) inhibitors, antidepressants, antiepileptics, etc. What makes a level of non-inferiority acceptable is whether the research question reflects a real need.4 If it does, independent trials, possibly supported by public funds, might be the reliable and affordable answer. There is no acceptable limit for hypotheses that serve only commercial purposes.

Secondly, equivalence or non-inferiority designs in clinical trials also reflect economic considerations. Marketing authorisation is easier to obtain when researchers have tested for and show equivalence rather than tested for and not shown superiority. In addition, trials that do not aim to show superiority require smaller sample sizes as they include in the “equivalence range” therapeutic differences that may well be clinically relevant if measured with the appropriate statistical precision.

Thirdly, patients participating in an equivalence or non-inferiority trial need to be informed clearly that the trial will not produce any real improvement in their health or the health of future patients. Patients should be told that there may be a risk of some harm with no hope of an advantage and that even if there were an advantage the trial design would not detect it. The patient should be told that the scope of the trial is purely commercial (see box).

A final reason for rejection of such trials can be shown by the example of the claimed equivalence found in some studies between tricyclic antidepressant agents and serotonin uptake inhibitors obtained under conditions that detected differences6 ranging from 12%7 to 43%.8 9 Clearly, looking for equivalence may look like an excuse for not seeking a difference. The advocates of equivalence trials claim that physicians need access to several drugs to select the best for each individual patient. Even if we disregard the fact that it is impossible meaningfully to select one drug rather than another from the array of ACE inhibitors, serotonin uptake inhibitors, non-steroidal anti-inflammatory drugs, or glucocorticoids, if the aim is to provide another opportunity for a patient who has not responded to a standard treatment why isn't the new drug studied in these resistant or non-responding patients?

Another frequent justification is that physicians may need drugs with the same effect but with different adverse reactions. This too can be rejected because the sample size in equivalence trials is usually too small to detect any reduction of infrequent side effects.

There are indeed cases where equivalence or non-inferiority trials are appropriate and do not pose ethical problems—for example, for testing a new drug with a more convenient route of administration (oral instead of injectable) or a less frequent dosing schedule. Equivalence can also be acceptable for public health reasons, such as a lower price or a reduced impact on the environment (such as chlorofluorocarbons-free preparations). Ethics committees must be constantly aware of these issues if they are to avoid patients being exploited and health resources wasted for clinical trials solely aimed at obtaining a slice of the pharmaceutical market.

Draft informed consent for an underpowered “equivalence trial”

Let us treat you with something that at best is the same as what you would have had before, but might also reduce—though this is unlikely—most of the advantages previously attained in your condition. It might even benefit you more than any current therapy, but, should that actually happen, we will not be able to prove it or let you know whether the new treatment may somehow bother or even harm you more than the standard one, as potential side effects may be too rare for us to be able to measure them in this study.

Comparators and doses

Research ethics committees must also look closely at the comparator and its doses and methods of administration to make sure the new drug does not start out with undue advantages. The comparator must be the best available and must be given together with the best available care for that indication. A few examples illustrate this. Clopidogrel has never been compared with the drug from which it originates, ticlopidine. Before it was compared with aspirin for the prophylaxis of thrombosis10 clopidogrel should have been shown to be more effective or safer, or both, than ticlopidine, otherwise there was no reason to compare it with drugs with different mechanisms of action. In addition, if it is not superior to ticlopidine why should the NHS pay for a drug that is much more expensive when generic ticlopidine is available? If clopidogrel is better why does ticlopidine remain on the market?

One of the first comparative trials tested risperidone at four doses compared with a single dose of haloperidol.11 It was therefore likely to identify a dose of risperidone that was better in terms of extrapyramidal effects. Subsequent studies have shown that risperidone is better than haloperidol only if the patient requires ≥ 12 mg of haloperidol, but there are no differences for patients who respond to < 12 mg a day.12 Sirolimus was superior to ciclosporin for the prevention of rejection after kidney transplantation.13 However, it was argued that the tested dose of ciclosporin was less than optimal, as established by the trough plasma concentrations.14 A recent review showed that when fluoxetine was tested as a new drug it had favourable effects in 70% of patients, whereas when it was used as a comparator it had favourable effects in only 58%.15 Fluoxetine was administered at high average doses (> 30 mg/day) in 43% of the original studies but in only 13% of the comparator studies. Clearly there are biases in clinical trials that often tend to favour new drugs.

Therapeutic end points

It is always important to look carefully at the variable selected to evaluate the benefit induced by a study drug. Ethics committees must require these end points to be meaningful for therapeutic action. Surrogate end points to predict clinical outcomes would help to avoid redundant research and the generalising of data from one member of a drug class to another. However, end points unrelated to morbidity and mortality cannot be regarded as satisfactory when effective drugs are already available.

For instance, two hypocholesteraemic agents, simvastatin and pravastatin, reduce mortality in patients with cardiovascular disease.16 17 Is it acceptable that other such drugs are marketed without similar evidence? Usually to avoid the need for this proof it is claimed that if the new drugs show a similar mechanism of action or biological effect, or both, they may share what is called a class effect. However, the case of cerivastatin has shown that certain adverse reactions—for example, rhabdomyolysis—may be more common with than with other statins.18 In this case the lack of data about the therapeutic activity made it impossible to assess the ratio of benefit to risk. Other examples are some COX 2 inhibitors, such as rofecoxib, which are less favourable than others in terms of cardiovascular effects,19 and atypical antipsychotics, such as olanzapine, which induces leucopenia less often than clozapine but is associated with a higher risk of suicide,20 diabetes,21 and weight gain,22 and impacts on mortality.23

Another tendency is the use of combined end points. Studies aimed at reducing cardiovascular morbidity often sum the rates of cardiovascular infarction, cerebrovascular events, and peripheral vascular occlusion because evaluating a single end point would require too many patients. However, the decision to use a combined end point may result in indications that do not reflect the drug's real value. For instance, in the CAPRIE study clopidogrel significantly reduced the combined end point,10 but close examination of the results shows that most of the effect was on the peripheral vascular end point24 rather than myocardial infarction or stroke. This involves the risk of accepting indications for which drugs with better activity are already available.


The problems outlined here raise important concerns that research ethics committees should consider an essential part of their ethical evaluation. We believe that most protocols contain elements that are not in the patient's interests (or may even go against them). Such protocols should not be approved because economic considerations have become more important than the real purpose of clinical trial. As the Declaration of Helsinki states: “The primary purpose of medical research involving human subjects is to improve prophylactic, diagnostic and therapeutic procedures.”


  • We thank Alessandro Liberati for fruitful discussion and useful advice, and Judy Baggott for helpful editorial assistance.

  • Contributors Each author substantially contributed to the conception, discussion, and debate of the issue raised. They equally contributed to the first draft, revised the paper critically for important intellectual content, and approved the final version. SG is guarantor.

  • Funding None.

  • Competing interests None declared.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
View Abstract

Log in

Log in through your institution


* For online subscription