The courts have always been America's institution of interpretation and last resort, fair and adept, most Americans believe, at presiding over criminal prosecutions and civil disputes.

But in recent years a disturbing trend seems to have emerged: using the decisive powers of the courts to determine what constitutes safe and effective treatment. Furthermore, judges are ordering insurance companies to pay for certain kinds of medical treatment, some of which are experimental.

Two recent cases raise questions about the capability of the courts to make decisions that traditionally have been made by professional, administrative and legislative bodies.

Several months ago in New Jersey, a U.S. District Court judge ordered Blue Cross and Blue Shield to pay for a bone marrow transplant for an 8-year-old girl suffering from a rare kidney cancer called Wilms' tumor. Payment had originally been denied by the insurance company because the procedure had been deemed "experimental."

While some treatments reside in the fuzzy area between experimental clinical trial and established therapy, bone marrow transplantation for Wilms' tumor is not one of them. The treatment, which costs about $135,000, is often associated with harmful side effects and has no scientific basis as established therapy in treating this disease.

In Northern Virginia several weeks later, the U.S. District Court in Alexandria was asked to make a similar judgment in another heart-rending case. This time, a young mother with breast cancer sought a bone marrow transplant as a last resort. Virginia Blue Cross and Blue Shield was ordered to pay for it. Once again, the treatment is experimental at best and hazardous at worst.

In both cases, the insurers faced a public-relations and managerial Catch-22. Denying treatment to a child and a young mother brought a public outcry against both insurers. At the same time, their own panels of scientists judged the technology to be experimental. The insurers may have hoped that the courts would spring the trap that had closed around them.

If the judges ordered reimbursement for expensive procedures, the insurers were off the hook with their directors and policyholders. If the courts declared the procedures to be experimental, their own initial judgment would have been vindicated. In either case, media coverage ensured that the public would know the outcome, and the buck was passed to two judges.

Court cases set precedents and can, in novel situations, substitute a judge's decision for the policymaking machinery usually reserved for Congress and state legislatures. Judges frequently feel uncomfortable in such cases.

Georgetown University's Program for Science and Law recently completed an extensive study on the future of the courts. It found that judges and court experts are worried about public expectations that the courts can or should solve problems that are too politically charged for legislatures or too financially sticky for government agencies.

In the case of experimental vs. established treatments, how are courts able to marshal the expertise that is the traditional province of doctors and academics? Experts who testify in court typically work for one side or the other -- hired guns in a high-stakes adversarial game. How can judges avail themselves of the impartial scientific knowledge needed to declare an expensive procedure safe and effective? How will the courts know when an order to pay for a procedure may be a sentence to a punishingly shortened life, or to one accompanied by serious medical complications?

In its recent report, the Federal Courts Study Committee, a congressionally mandated, blue-ribbon panel that examined the federal court system over an 18-month period, recommended that the management of scientific and technically complex matters such as these be incorporated into a manual to guide judges. We support such an approach. The march of science and the explosion of technical information warrant it. But procedural guidelines alone won't suffice in the years ahead.

The courts need two substantive capabilities. The first is direct access to a medical technology assessment process they trust as impartial, thorough and reflective of the state of the art. "Technology assessment" is the formal evaluation of a device, drug or procedure used in health care to examine its safety and effectiveness as well as other ramifications.

In addition, judges need regular access to experts who don't have vested interests and to whom they can defer when making a decision about what constitutes accepted therapy or experimental treatment. A number of excellent technology assessment programs have been established in this country, but if the courts find that none of them is adequate, it is time to create a new mechanism directly responsible to the judiciary.

The courts increasingly will be called upon to make life and death decisions from a civil, rather than a criminal, perspective. If they accept this role, it is in the public's interest that they better equip themselves to render these awesome decisions.Franklin M. Zweig is director of the Program for Science and Law, Department of Community and Family Medicine, and adjunct professor of law at Georgetown University. Seymour Perry is interim chairman of the Department of Community and Family Medicine and director of the Program on Technology and Health Care at Georgetown University School of Medicine. Second Opinion is a forum for points of view on health policy issues.