Medical science has made enormous strides since the release of the first national Dietary Guidelines in 1980, but many would agree that the quality of the American diet appears inversely related to these health gains. We've added years to our lifespan and provided the safest, highest quality foods possible yet, as a nation, the quality of our diet has deteriorated. The Secretaries of Health and Human Services and of Agriculture will soon name the next DGAC to define the science base for the 2010 Guidelines. This is the place to fix the problem. We need to establish the new 2010 Guidelines as worthy of the trust Americans hope to place in them as an authoritative source of information about their food choices.

The Secretaries should consider carefully the critical importance of the selection criteria for Committee members. With obesity at historic levels and childhood obesity a near epidemic with grave long-term consequences for our nation, the need for policy guided by expert scientists is clear. No one could deny that the first six DGACs were composed of prominent medical and nutrition experts. There can be no quarrel with the professional, subject matter qualifications of past DGAC members. It's the paradigm that needs changing.

Committees of subject matter experts produce reports with expert opinion. That sounds better than it really is. In the hierarchy of "evidence-based" medicine, expert opinion is the lowest level of evidence. Rigorous data analysis trumps even well-informed opinion. To sort out public confusion and establish consensus authority, we need to move higher on the evidence-based hierarchy. We must do better for the nation. Evidence-based decision-making focuses less on the experts and more on the evidence. While as good they could be, because past DGACs have not followed the best discipline, their reports cannot claim the mantle of evidence-based reviews. We need to change the DGAC process, not just the people on the Committee.

Using a process like the one developed in the 1980s by the Cochrane Collaboration , inventors of "evidence-based medicine," will allow this new DGAC to take the next step in the process and set the standard and grade the evidence before considering the policy analysis. We need this different expertise on the DGAC in order to make the Guidelines reflect the science and become most relevant to Americans' health.

The federal government endorses an evidence-based approach to health policy and the U.S. Preventive Services Task Force (USPSTF) is the model for how the DGAC could adopt the discipline of evidence-based decision-making. Supported by the HHS Center for Outcomes and Evidence and a contracted Evidence-based Practice Center in Oregon which conducts systematic reviews of the evidence, the USPSF makes its recommendations on the basis of explicit criteria. The USPSTF reviews the evidence, estimates the magnitude of benefits and harms for each intervention, reaches consensus about the net benefit for each intervention, and issues a recommendation - from "A" (strongly recommends) to "I" (insufficient evidence to recommend for or against).

The USPSTF process would be the most appropriate and effective model for the DGAC. An evidence-based review will require an evidence-based process; this requires changing the concept of the DGAC which up until now has been compiling expert opinion instead of conducting an evidence-based review.

In a courtroom, judges rely on subject matter experts: witnesses attest to their observations and "expert witnesses" offer their professional opinion. Judges are not subject matter experts; they are process experts. They know what observations and opinions to admit into evidence. They discipline the process. The DGAC has been acting as an "expert witness" instead of a judge. We need a DGAC composed of "judges" - experts in the process of evidence-based decision-making. We need "judges" who have a proven dedication to dispassionate review of the evidence. And we need their report to reflect their conclusions about the quality of the evidence before the policy conclusions and recommendations.