Are Diabetes Rates Really Holding Steady?

Yes, according to this article by Gina Kolata in today's New York Times. But a close review of the data contained in the study that informed the article yields a less sanguine interpretation.

The study compared the changing rate of diabetes, both diagnosed and undiagnosed, and pre-diabetes between the early 1990s and the first years of this decade. The study relied on data drawn from the well-regarded National Health and Nutrition Examination Surveys (NHANES).

While diagnosed diabetes rates increased significantly from 5.4 to 6.5 percent of the total population (12.8 to 15.8 percent for those over 65), the undiagnosed diabetes rate and ratio of people with impaired fasting glucose remained unchanged. The easy conclusion to draw -- and both the study's authors and Kolata drew it -- was that the health care system is doing a better job at identifying and treating diabetics, but the obesity epidemic hasn't really had much impact on the overall rate.

But I came away from the data with a different and more alarming conclusion. The total population with diabetes (whether diagnosed or undiagnosed) and impaired glucose tolerance reached 35.3 percent or 73.3 million Americans by 2002. While that was deemed not "statistically significantly" greater than the 33.7 percent with those conditions in the early 1990s, in crude terms, it amounted to at least 3 million more people at risk of losing eyesight, limbs, heart attacks, strokes and the other debilitating co-morbidities of diabetes.

Moreover, the lack of statistical significance was wholly an artifact of statistical manipulations to the crude data to adjust for changing demographics of the population. If one just looks at the raw numbers, the rate of diabetes, both diagnosed and undiagnosed, leaped from 7.8 percent of the total population in the earlier period to 9.3 percent in the early part of this decade. And whether we look at the raw numbers or the statistically manipulated numbers, the over-65 crowd with diabetes jumped more than three full percentage points to well over 21 percent.

Given the amount of change seen in such a short period of time (about a decade), you'd think the researchers, who hailed from the National Institute of Diabetes and Digestive and Kidney Diseases, a unit of the National Institutes of Health, would have put a different spin on the overall data, even while recognizing its lack of statistical significance. The trend certainly seems to be clear.

Indeed, if I were writing up the study, I would have noted that while the health care system is doing a better job by identifying people with diabetes, there is, alas, something about the way we live today that is fully replenishing the ranks of those who need to be diagnosed.