The overarching mission of this respected blue ribbon panel was to suggest how training could be "enhanced" to "better prepare" medical subspecialists for their "future roles." The premise for the Task Force's deliberation was that in the face of a widely accepted surplus of many medical subspecialists [5, 6] and almost certain downsizing of residency and fellowship programs [2, 3, 7], the quality of education must be protected and the opportunity for constructive change should be seized.

The main recommendation of the Task Force is a proposal for a bifurcated educational pathway of either subspecialty "clinician" or subspecialty "investigator," with the latter track having two options: basic scientist and clinical investigator. Given the new era of managed care and the focus on enrolled groups, the Task Force correctly suggests that clinical research will increasingly emphasize epidemiologic analyses and cost-effectiveness and outcomes research. For the foreseeable future, this class of subspecialist-health services researcher will be in great demand. Furthermore, the continued survival and growth of the individual medical subspecialties will rest, in part, on the abilities of this cadre of interdisciplinary scientists. Because more technologies and providers exist than society can (or will) pay for, these investigators will be asked to provide evidence to health plans and policymakers on the value of services provided by medical specialists.

The Task Force suggests that every training program in the United States should provide both educational pathways and that, "on occasion," concentrating resources within fewer programs may be cost-effective. One can only assume that this suggestion was couched in such terms because of the Board's limited aegis or the ever-delicate nature of the political coalitions that have come to characterize internal medicine's organizational factions. The Task Force's exhortation is neither realistic nor desirable. Although it is true that all clinicians would benefit from exposure to research and that all subspecialty investigators must first become rounded clinicians, the Task Force's secondary suggestion of multi-institution consortia is more on the mark. Given that many current clinical training programs can be considered only marginal, can these programs also be expected to offer a comprehensive twopronged research pathway?

A related issue that also appears to have been side-stepped is a delineation of the exact role the Board (and other internal medicine groups) should have in the downsizing process. As the total number of training slots decreases, exactly how should accreditation be used to ensure that only the best programs remain within each subspecialty? The report suggests that "the proportion of subspecialty clinicians should be reduced and the proportion of investigators increased"; however, the Task Force did not address how extensive the reduction should be or who should govern and fund this process. It is no easy task to answer either question.

A second major recommendation cited on the Task Force's list is promotion of the concept of the medical subspecialist as a "principal care" provider or the team leader for patients with complex illnesses or specialized needs [8, 9]. For some patients, this expanded specialist role will undoubtedly improve outcome or quality of life; for other patients, however, this will not be the case. This debate must be guided by patient-centered, empiric cost–benefit research, not by the more limited economic self-interest of providers or payers. Furthermore, this deliberation must not occur only within medical subspecialty circles. General internists, non-internal medicine physician specialty groups (such as family practice), and nonphysician providers (such as nurse-practitioners and physician assistants) must also be at the table.

The American Board of Internal Medicine should be commended for initiating what must be viewed as the beginning of a long and difficult process. As science and society evolve, so too should educational institutions; this process must be continuous. As this and other policies affecting the training of 21st century clinicians and investigators are established, we must constantly remember that the needs of the population must remain paramount. The ultimate goal should be the maximization of quality of life for all, balanced against the United States' very real resource limitations. Within this frame of reference, the track record of our medical education system in the last several decades has been less than stellar. We lost sight of our mission, and precious resources were misdirected toward inappropriately specialized technologies and personnel. These mistakes must not be repeated, and other mistakes must be avoided. We must guard against "policy overshoot," in which, because of cost cutting and antispecialist sentiment, we end up with a less specialized workforce than we require. We must get it right. Too much is at stake.