• Chart: The US isn’t good for immigrants when it comes to diabetes

    This is the percent of people ages 18 to 64 in the United States ever diagnosed with diabetes, by region of origin and ethnicity or race, 2009-2011 (page 54):

    Percent with diabetesImmigrants from a variety of places have a very low incidence of diabetes for up to 10 years. Those who live in the US longer, however, fare much worse. One of two things is going on here. The first could be that people arrive here relatively healthy, and then America is so bad for you that they develop diabetes in large numbers. The second is that they already have diabetes, but it’s undiagnosed, and our care of them is so bad that we’re not picking up on it for at least ten years.



    • There’s a lot of literature out there supporting your first explanation, particularly for low SES immigrants. A starting place is the recent literature review on the subject: “Diet or Exercise? Evidence from Body Mass Index of U.S. Immigrants”:


      Your second explanation needs clarification. We are either diagnosing them earlier than they would’ve been diagnosed in their birth countries (which, again, is reasonable for low SES immigrants from Mexico, India, etc.). Or, our care is so bad that we’re enabling a faster disease progression. I find that unlikely, but your wording makes it sound like that’s your intention. Which are you suggesting?

      • I interpreted the statement to mean that it is possible that these individuals are living in the US for a long time with undiagnosed diabetes. I don’t see how that would require a faster disease progression.

    • It is difficult to determine diabetes duration at diagnosis. It would be interesting to know the prevalence of diabetes complications like retinopathy or nephropathy at diagnosis. A high prevalence could indicate a long interval of undiagnosed diabetes.

    • While I would agree with one, I would note that if people are most likely to immigrate when they are younger then the population 18-65 that has stayed in the U.S.for 10 years is likely to be older than the recent immigrants and so is not a completely comparable population. Now if they had compared the 18-25, 26-34 groups it would be more interesting.

    • I agree with sort_of_knowledgable … the populations aren’t really that similar and a better analysis (though it may not exist) would track a control group of people who never immigrated and stayed in their home country. Age is a key factor here – you could then control for the age difference if you saw, for instance, the two-bar result for Mexicans who stayed in Mexico of roughly the age of the ‘recent arrivals’ and ‘long stays’. Difficult to parse out, though…