An Increase in Diagnoses May Not Mean a Higher Rate of the Disease, a Survey Shows

By GINA KOLATA
The number of diagnosed cases of diabetes increased by 61 percent from 1991 to 2001, according to the federal Centers for Disease Control and Prevention. That went along with a 74 percent increase in obesity, the agency noted, “reflecting the strong correlation between obesity and the development of diabetes.”
But those numbers may or may not reveal that the actual number of people with diabetes has exploded. It may just be that more people are learning they have the disease, not that the number of those with it is increasing.
To get a better idea of whether the disease is striking more people or whether more people who have the disease are receiving diagnoses, statisticians have turned to another set of federal data. It is from the National Health and Nutrition Examination Survey, a periodic survey of a representative group of Americans that not only asks whether people have been told they have diabetes but that also includes blood tests to find undiagnosed cases.
In a paper published last year in Diabetes Care, federal scientists used those data to ask what was the total proportion of the population with diabetes — diagnosed and undiagnosed.
Their surprising conclusion, said Katherine M. Flegal, an author of the paper and an epidemiologist at the National Center for Health Statistics, was that the overall age-adjusted proportion of the population that has diabetes had not really changed from 1988 to 2002, the most recent year for which federal data are available.
In two subgroups, though — men and non-Hispanic blacks — there was a small but statistically significant increase in the diabetes rate. In general, blacks, Hispanics and Asians are more likely to develop diabetes than whites, and regions with many immigrants of those racial and ethnic groups may experience a surge in the population of people diagnosed with diabetes. But that, of course, does not reveal whether the nation as a whole is experiencing a sharply increasing rate of diabetes.
Still, says James Smith, a senior economist at the Rand Corporation, who independently analyzed data on men from the national survey, the data are puzzling. Why would there be a significant increase in the diabetes prevalence in men but not in women, since the changes in underlying risk factors like obesity are similar? And why is there no increase over all?
Dr. Smith suspects that the prevalence data were distorted by data on diabetes in women that may not be as reliable as those involving men. If he is correct, he said, it may be because the surveys in different years were not consistent in dealing with gestational diabetes, an often temporary condition that can occur during pregnancy. When he takes that out of the data, he sees similar rises for men and women.
Dr. Smith said he did not want to diminish the significance of the diabetes problem in this country. Indeed, he said, the problem is increasing. But he said there is little question in his mind that the confusion between the number of people who have been told they have diabetes and the actual number of people who have the disease has led to an exaggerated perception of how rapidly the disease is spreading.