Has there really been, as the CDC reports, a huge decline in diabetes complications? | Bleader

Has there really been, as the CDC reports, a huge decline in diabetes complications?

by

comment

Apples and oranges?
  • CRSPIX/Shutterstock

Here's the good news according to a study from the Centers for Disease Control and Prevention, published this week in the New England Journal of Medicine: if you have diabetes, you're 67.8 percent less likely to die of a heart attack than you would have been in 1990.

You're also more than 50 percent less likely to have a stroke or an amputation.

Here's the bad news: you're about 300 percent as likely to have diabetes.

Why? Maybe because the criteria for diagnosis changed in 1997, lowering the blood sugar threshold and bringing hoards of healthier people (less likely to develop complications) into the population of diabetics. There were 6.5 million adults with diagnosed diabetes in the U.S. in 1990, and 20.7 million in 2010.

I couldn't find that explanation in the study itself ("Changes in Diabetes-Related Complications in the United States, 1990-2010"), though. The only allusion to it is an oblique admission that results "could also have been partially influenced by changes in the characteristics of the population, given differences in diagnostic criteria . . ."

Instead, the authors suggest the astounding decline in complications is likely due to better care and management (including the use of prescription drugs).

"We can't discount [the change in diagnostic criteria] as a contributor," lead author Edward Gregg told me today, after I dialed up the CDC to ask. "But we do not think it was a major contributor."

"These decreases in complications are of such great magnitude they couldn't have been accounted for by that," Gregg said. "And we continue to see the decline in the complications all the way through the 2000s, so that would imply that the change in the diagnostic criteria was not a big factor."

He directed me to a table published with the study that tracks those declines. But it shows a leveling off between 2005 and 2010.

And I'm pretty sure I learned something about apples and oranges in Statistics 101. Like a change in the population measured is going to corrupt the results.

Add a comment