During a recent weekly staff conference at Massachusetts General Hospital, the custom at another hospital -- paying a $350 finder's fee to other physicians for each patient who enrolled in a new study -- was mentioned.

At first glance, there may not appear to be much wrong with finder's fees. Some might consider them a reasonable solution to the problem of recruiting patients for clinical trials. But the more I thought about it, the less acceptable the practice seemed, particularly in medicine.

As with many activities, finder's fees occupy a spectrum of acceptability. While there is probably little to be concerned about when a health club offers members a discount for recruiting new members, it is unethical -- not to mention illegal in many cases -- for a business executive to offer a fee to a government official for "finding" a government contract for his or her company.

Although I'm sympathetic to the difficulties experienced by investigators who need to recruit adequate numbers of patients to their studies in order to reach accurate conclusions, finder's fees for research subjects are too close for comfort to the practice of fee-splitting.

Fee-splitting, long prohibited by major medical organizations, occurs when a physician receives money for referring a patient to another doctor. It dates back to a time before the reform of medical education in the early 1900s, when many doctors were not well trained. As a result, patients were frequently sent to consultants because a fee would be given to the referring physician, not because the consultant could offer high quality care. It was recognized then that patients may not be able to fully evaluate medical advice and often simply accepted their physician's advice based on trust alone.

Physicians have long recognized that they must go to special lengths to minimize the potential, apparent and/or real conflicts of interest inherent in many aspects of medical practice. Because it is important to ensure that medical advice is as unencumbered by self-interest as possible, medical fee-splitting is prohibited today, even when patients are referred to capable and competent consultants.

For the same reason, finder's fees should not be offered for the referral of research subjects. Physicians and patients should never have to wonder if professional judgment is being compromised by the possibility of personal gain of whatever magnitude.

While the risk of harm to a patient referred to a carefully conducted trial because of a finder's fee is probably insignificant, efforts must be taken to strengthen the public's trust and participation in clinical research, rather than risk harming it in any way.

We should not assert that clinical trials provide uniform, meticulous care designed by experts in the field and then offer bounties for referring patients. If the circumstances are appropriate for suggesting to a patient that he or she consider participation in a clinical trial, then the referral should be made.

It's not difficult to understand why physicians might not automatically think to refer their patients to clinical trials. First, they would lose, at least temporarily, the personal contact with the patient so central to enjoying the practice of medicine.

Furthermore, they might lose contact with the patient forever, should the investigator take over the patient's care when the study ends. This might hurt the physician personally, financially or professionally, and might be particularly important in regions where there is significant competition among doctors for patients.

Some physicians might not refer patients because of bias against clinical researchers. There is an underlying tension between investigators and practitioners in many communities, which may have complex roots.

In some cases, this arises because investigators have not maintained good lines of communication with the referring physician or failed to be sure that the patient returned to the referring physician at the completion of a study.

In other cases, it may reflect attitudes conveyed to young physicians in training. Unless special efforts are made during their training, residents may enter practice without having been exposed to careful investigators and without a sense of responsibility for continuing the advancement of medical knowledge, either by participating in clinical research or by referring appropriate patients to those who do.

Programs that foster mutually beneficial interactions between practitioners and researchers, such as those initiated by the National Cancer Institute, are valuable in overcoming the barriers that separate these distinct groups of physicians. The needs of clinical research must be explained to medical professionals as well as the public.

With medical progress comes complexity; more combinations of tests and treatments must be scientifically compared to see if they really are superior to earlier or simpler treatments.

Since it is certainly not ethical to experiment on people without their knowledge or consent, recruitment of patients to clinical trials becomes a societal concern, not just one restricted to researchers. Good physicians, and increasing numbers of policymakers, are aware that the standards of medical practice are best determined by carefully performed studies in different groups of patients.

Patients and physicians must have realistic views of what modern medicine can and can't do. They must understand what is involved in achieving medical progress. Patients must contribute to the continual process of evaluation that characterizes and informs the best medical care.

If patients and physicians are interested in ensuring that everyday patient care is the best possible, they will seek out opportunities to promote clinical research, and finder's fees will no longer be needed.

Stuart E. Lind is a hematologist at Massachusetts General Hospital.

Second Opinion is a forum for points of view on health policy issues.