UCLA researchers mined patients' electronic health records to uncover some surprising risk factors for Type 2 diabetes.
First, the good news: being prone to migraines reduces your risk for diabetes by the same amount as being 29 years younger. But having a history of viral infections and chlamydia increases your risk for diabetes by almost as much as having a high body mass index.
Their research was published today in the Journal of Biomedical Informatics. It's unclear why these other conditions increase people’s risk for diabetes.
The findings are based on electronic health records of 9,948 people from 50 states. The records included vital signs, prescription medications and reported ailments - but not patient names. The researchers used half of the records to create an algorithm that would predict the likeliness of someone having diabetes; they then tested this algorithm on the other half.
Lead author Ariana Anderson, an assistant research professor and statistician at UCLA's Semel Institute for Neuroscience and Human Behavior, said the findings show the value of keeping electronic health records - a hopeful message for doctors who find it laborious and time-consuming.
"All of this information and the work that they are doing isn't going down the drain," Anderson said.
Traditionally, doctors have determined whom to screen for diabetes based on known risk factors, like age, gender, BMI, hypertension and smoking. Now, Anderson said, the data can also help diagnose diabetes and other conditions, before people develop complications.
"There is still enough predictive power left in these little diagnoses, such as migraines or viral and chlamydial infections, that you can actually use these to help predict the disease risk, on top of everything else you already know," she said.
If this algorithm were used nationally, researchers calculate that it would identify 400,000 people who have not yet been diagnosed with diabetes.
Beyond diabetes, electronic health records could also be powerful tools in screening for rare diseases like lupus, Anderson says.
"It's certainly not going to be cost-effective to do screening for everyone for lupus, but we can perhaps create sufficient models that we might be able to identify people who are more likely to have more rare diseases than diabetes," she said.