Assessment of Scientific Design

IRBs are established to safeguard the welfare of the subjects of research and not to provide rigorous peer review of the scientific merits of a proposed study. Nevertheless, the norm of sound scientific design, based as it is on the ethical principles of beneficence and respect for persons, and the statement of policy that flows from it, require that IRBs review the scientific basis for proposed clinical trials and assess the scientific and statistical design of the trial. Among the information needed for this review and assessment are (a) the results of animal studies and of previous clinical studies or experiences in humans, (b) whether there are similar studies currently underway elsewhere, (c) the scientific rationale for the study being proposed, and (d) the statistical basis for constructing the trial.

The manner in which an IRB processes relevant information is illustrated in the following example. A cardiologist wishes to evaluate a new antiarrhythmic agent. An IRB reviewing the proposed study should ask whether the drug has been previously tested in animals or humans and, if so, under what circumstances and with what results. The IRB should also inquire whether similar trials are presently being conducted elsewhere and whether all patients in the proposed trial will receive the drug or patients will be randomly allocated to receive either the drug being evaluated or current standard therapy. If the trial is designed as a randomized clinical trial, the IRB should satisfy itself that it is constructed in a manner allowing for the accrual of sufficient numbers of patients and allocates subjects between the proposed treatment and the standard therapy in a manner that allows the investigators to draw conclusions regarding the relative effectiveness of the two treatments.

A thorough analysis of these matters generally requires substantial medical and statistical expertise. It is to be remembered that the federal regulations require that the membership of IRBs not be drawn exclusively from one professional group, e.g., physician-investigators. Rather, IRB membership should include representatives of a variety of disciplines and professions in order to reflect more appropriately the values of society at large. Given their composition, how can IRBs exercise their responsibilities for review of research design?

It is suggested that one of two approaches may be adopted by those IRBs lacking the requisite expertise to assess research design. One approach, which may be particularly suitable for IRBs in community hospitals, is merely to accept the information included in the project proposal as providing an adequate basis for justifying the study and a suitable design for achieving its purposes. This approach might be acceptable in the case of a multi-institutional clinical trial conducted under the sponsorship of one of the National Institutes of Health. Such trials have had the benefit of outside review groups such as the National Heart, Lung, and Blood Institute and the FDA during their development and prior to their submission to IRBs for approval.

In the development of the proposal, the investigators responsible for the study would have addressed the specific issues of scientific rationale and statistical design. These individuals or others with whom they might have consulted would presumably have the training and experience in the disease being treated and in statistical methods. The design of the proposed trial would be reviewed by personnel from the sponsoring Institute and, in the case of trials involving investigational new drugs, by the FDA. The proposal submitted to the IRB would have had the benefit of an analysis with respect to research design substantially more extensive and more sophisticated than that which could be undertaken by most IRBs. Under these circumstances, it would be appropriate for an IRB to take cognizance of the prior review and not proceed with an independent analysis of scientific design. Indeed, it could be argued that an additional review of scientific design by an IRB lacking the expertise available to a scientific peer review group would not contribute further to safeguarding the welfare of the subjects.

This approach would not be acceptable in the case of a clinical trial proposed by an individual investigator that had not been exposed to review by outside agencies. In the case of such proposals, IRBs have a quasi-peer review role in assessing the design of the studies. In these instances, IRBs have a relatively more significant function in safeguarding the welfare of potential subjects than when proposals have had prior outside peer review. It is unjustified to omit this assessment simply because of the practical problems in carrying it out. Additionally, to do so would result in the IRB being out of compliance with the spirit and the letter of the federal regulations.

To carry out the review of research design of investigator-initiated proposals, IRBs can adopt an alternative approach, that of soliciting the opinions of outside consultants. This approach can be utilized by university hospitals as well as by community hospitals and is applicable to the review of multi-institutional trials as well as those proposed by individual investigators. The consultants should be individuals familiar with the medical issues in question and with statistical analysis. They would determine whether suitable information exists to justify the proposed trial and whether the trial as designed would achieve its intended purposes.

The outside reviews would not be intended necessarily to determine whether the proposed trial is the best one that can be done. They could well lead, however, to improvements in the study design that would benefit both subjects and investigators. The use of consultants would thereby enable IRBs to compensate for limitations of expertise on the part of their members and allow them to meet their responsibilities for assuring satisfactory research design.

0 0

Post a comment