President Obama used his considerable powers of persuasion to try to sell his health care package in a nationally televised press conference this week. But Americans are growing skeptical -- and for good reason. The gargantuan new bureaucracy Obamacare envisions would not only be inefficient and expensive but could give birth to a new racial spoils system.

Among the provisions in the thousand-page House version are special set-asides aimed at training "underrepresented" minorities in health care professions. The idea is that some minority groups -- but not all -- will be better served if their doctors share their racial and ethnic background. It's an idea that has been floating around for years.

In 2002, the Institute of Medicine released a study entitled "Unequal Treatment: Confronting Racial and Ethnic Disparities in Health Care" that sparked a flurry of accusations that minority patients, especially African-Americans, receive bad health care because their doctors were biased. The study said that "(s)ome evidence suggests that bias, prejudice and stereotyping on the part of health care providers may contribute to differences in care." But as Dr. Sally Satel, a highly respected physician and author, observed at the time, the "evidence" in the study was thin. "'Some,' 'suggests' and 'may,'" she wrote, "are all the kinds of words authors use when the data are flimsy and reputations are at stake."

There is no question that African-Americans, on average, die younger and have poorer health than whites. What is less clear is why that is the case. Socio-economic class and behavior both play an important role. Homicide is the leading cause of death for young black males between the ages of 15-24, for example. Obesity, drug and alcohol use, and other behavioral factors play an important role in determining overall health. But will insisting on preferences for African-American students applying to medical school admission improve health care for blacks? Not likely.

A 2006 study by my Center for Equal Opportunity on preferential admission practices at the University of Michigan School of Medicine showed that admitting black students with lower grades and test scores is a bad idea -- especially for the patients who might be treated by these doctors. Black students admitted by UMMS had substantially lower test scores and undergraduate science grade point averages than all other groups admitted.

Indeed, in the four years of data CEO analyzed, 11,647 white, Asian, and Hispanic applicants were rejected by UMMS even though they had better grades and test scores than the average black student admitted. And UMMS isn't alone in using race to determine who gets in. CEO has studied preferences in medical school admissions at more than a half-dozen medical schools, including the University of Maryland, the University of Washington, and the State University of New York Brooklyn, all with the same overall results.

African-American med students who are admitted despite having lower grades or test scores than their white or Asian peers are less likely to pass medical licensing exams -- or, even if they pass, are more likely to perform poorly on them. This hurts everyone: the better-qualified students who are passed over to admit those who will eventually fail to become doctors, and the patients who may end up treated by doctors who are less well prepared. If the idea is to get more black doctors to treat black patients, is it really going to improve health care for African-Americans if those doctors perform worse on medical licensing exams?

But Obamacare will push more institutions to adopt racial preferences by giving preference to those that have, in the words of the Democrat House legislation, a "demonstrated record" of "training individuals who are from underrepresented minority groups or disadvantaged backgrounds." And notice the term "underrepresented minorities." They may as well have put up a sign "Asians need not apply."

Other provisions in the Democrats' bill would provide for "maintaining, collecting and presenting federal data on race and ethnicity," in order to "facilitate and coordinate identification and monitoring ... of health disparities to inform program and policy efforts to reduce such disparities." We've seen these efforts before in the context of employment and education. Their end result is always a form of bean-counting that leads to racial quotas -- which is bad medicine and won't improve health care for anyone.