Thousands of Latino patients stream though the East Los Angeles practice of Dr. Hector Flores and his partners each year.
The older ones go to the family practice with arthritis and hypertension, the younger ones with diabetes and asthma.
What surprises Flores, however, is not how sick they are, it is how sick they are not.
Overall, Flores said, his patients are much healthier than one would expect given their low levels of income and education, factors epidemiologists long have known are linked to poor health.
"You can predict in the African American population, for example, a high infant mortality rate," he said recently, "so we would think a [similarly] poor minority would have the same health outcomes.
"But they don't. They're not there," he said, referring to outcomes among Latinos.
Why Latinos aren't sicker -- a phenomenon known to health experts as the Latino paradox -- is puzzling to public health experts, given the link between disadvantage and high disease and mortality rates.
In overall mortality rates and infant mortality rates, two standard measures of a population's health, Latinos' numbers approach and sometimes surpass those of whites.
In Los Angeles County in 2003, the age-adjusted mortality rate for Latinos was 535 per 100,000, 33% less than for non-Hispanic whites and 52% less than for blacks, according to the most recent data from the county's Department of Public Health.
Nationally that year, Latinos' mortality rate was 621, 25% less than whites' and 43% less than blacks', according to National Vital Statistics Reports, published by the Centers for Disease Control and Prevention.
Latinos' infant mortality rates reflect a similar pattern. Locally, the rate was 5.2 per 100,000 in 2003, 16% higher than whites' and 57% less than blacks'. The national rate was 5.7, about the same as whites' and 58% less than blacks'.
"It violates one of the most predictable patterns we see in most areas of the world and for most diseases," said Dr. Paul Simon, chief epidemiologist for the county's Public Health Department.
"The question is, 'What is the Latino population doing right?' "
The reasons for the paradox are a matter of some debate. Some scholars attribute it to immigration, which may draw selectively from the ranks of the hale and hardy.
Another possibility is that many immigrants return to their home countries when seriously ill, skewing mortality statistics in this country.
But increasingly, researchers are suggesting that such factors as diet, lifestyle choices and strong social support networks are the key to Latinos' better-than-expected health.
"They're not taking some secret Aztec herb they didn't tell you about," said David Hayes-Bautista, an early observer of the Latino paradox who directs UCLA Medical School's Center for the Study of Latino Health and Culture.
It's worth figuring out what is making the difference, he added, because "we could all be better off for it."
As the immigration debate heats up and the cost of healthcare soars, the phenomenon is attracting attention from social scientists and public health officials. It was first noticed, though, in the 1970s and 1980s by researchers looking at infant and overall mortality rates in Texas and California.
Scholars tended to view the findings as wrong or anomalous, assuming that Latinos' relatively disadvantaged socioeconomic status put their health status more on par with blacks'.
As data accumulated, covering broader swaths of the country and longer periods, skepticism turned to curiosity.
Kyriakos Markides, who in 1986 coined the term "Hispanic epidemiological paradox," described scholars' shift in recent years as remarkable.
"Nobody talked about it then," said Markides, a professor of socio-medical sciences at the University of Texas Medical Branch, speaking of a generation ago. "People just ignored the data or assumed that disadvantaged populations have high mortality. Now, it's the leading theme in the health of the Hispanic population in the United States."
The paradox remains controversial, in part because its first-blush message -- that Latinos are less ill than expected -- might lead people to believe that the group doesn't need scarce healthcare dollars.
Hayes-Bautista recalled an incident in the 1990s when a colleague he had just briefed on the paradox asked him to keep the data to himself, saying she feared services for Latinos could be cut.
"I guess it's the triumph of -- I hate to say -- of ideology over data," he said.
In addition, some doctors and other medical professionals aren't familiar with the concept, which is a broad statistical phenomenon, not necessarily true in individual cases or practices.
Even some who are aware of the phenomenon question its significance, saying they still must focus on treating individuals.