Abstract

I. Introductory Comments   The Human Genome Project will be completed within 2 years, and “targeted†sequence data from the most promising sections of the genome will be released even sooner. Based on this wealth of information, at least 400 new genetic tests will become available within the next decade. The blending of microelectronic and genetic technology will make the “genetic report card†an affordable and routine part of medical care. The implicit assumption driving much of this push for genetic testing is that information is inherently good and patients should have the right to any information about themselves they desire. However, the informational content of genetic tests for most conditions is not known and may in fact be very low. Under such circumstances, it seems prudent to treat access to genetic testing the same way we currently treat access to other new medical procedures and drugs – that is, access is afforded only 1) in the context of a clinical trial or  2) after the safety and effectiveness of the treatment has been established. The fact is, genetic “information†of uncertain quality is not innocous. II. The Difficulty with Genetic Testing   Everyone is familiar with one difficulty posed by genetic testing: the possibility that a test may show either a “false positive†or a “false negativeâ€.  This is the problem of the analytic accuracy of testing. However, as our analytic techniques become more sophisticated, the incidence of such errors will steadily decline. If nothing else, administration of multiple tests is often (though not always) a simple and effective way of boosting analytic accuracy.   There is, however, a much more serious accuracy problem which has been largely overlooked in the literature..