establish the performance characteristics of the device [emphasis added]" (Tsakeris and Yoder, 1992, p. 12).
The submission of protocols to IRBs for investigations covering such devices may overwhelm some IRBs. At some large academic centers, IRBs are appointing subcommittees to deal with new genetic investigations. Most IRBs will not have experts in genetics or individuals knowledgeable about ethical problems in genetic testing. The National Human Genome Research Center at the University of Iowa has created a unit to advise IRBs and other groups on ethical, legal, and social issues arising in genetics research.
At some point in their development, genetic tests emerging from research laboratories in academic medical centers are offered as a clinical service. Research laboratories sometimes perform these tests at their convenience, with delays of months in reporting results (Hoffman, 1991; Klinger, 1992). If no manufacturers are interested in commercializing the tests as in vitro diagnostic devices, the research laboratory continues to provide them as a service. In many instances, academic research laboratories do not comply with investigational-use device requirements of FDA, despite the fact that performance of the test for clinical purposes requires them to do so.
FDA is aware of the problem of the use of medical devices that it has not reviewed for marketing and that are not in compliance with its regulations regarding investigational devices. One example is the use of the MSAFP test (which has FDA approval for detection of increased risk of fetal neural tube defects) for detection of increased risk of Down syndrome in the fetus (for which the test is not approved). FDA has proposed to deal with such problems by compiling an "accommodation list" of such tests that it believes to be essential to clinical care. Manufacturers or other sponsors of the tests on this list would be allowed an additional 30 months to collect data for premarket approval applications. Tests not on the list can no longer be legally provided unless they immediately comply with FDA requirements (FDA, 1992). ACMG has pointed out to FDA that the number of genetic tests in this category is very large and that the elimination of tests that are not listed will cause serious problems to patients with, or at risk for, genetic disorders (M. Watson and M. Cohen, letter to F. Yoder, FDA, August 18, 1992). Because many new genetic tests will be for rare diseases, the collection of adequate data to establish safety and effectiveness will take a long time, probably longer than the 30-month time limit given by FDA for devices on an accommodation list. For tests performed only occasionally, "provisional premarket approval" should be given (see below). Once granted, this will permit laboratories to charge