
How a Novel Algorithm Can Improve the Prognosis for Type 2 Diabetes
Transforming patient care, one algorithm at a time
Dimitris Bertsimas was a young boy in Athens, Greece, when his mother was diagnosed with Type 2 diabetes. He was already familiar with the disease—a chronic, hereditary condition that causes blood glucose (sugar) levels to rise higher than normal—because his grandfather had died of complications related to it. His mother’s sister, who lived only a few streets away, also suffered from the illness.
Even as a child, Bertsimas recalls being puzzled by the fact that his mother and aunt received such very different treatment from their respective physicians. His mother never took insulin, a hormone that regulates blood sugar levels; instead, she ate a restricted diet and took other oral drugs. His aunt, meanwhile, took several injections of insulin each day and dealt with many more serious side effects.
“Back then, there was no way to provide targeted treatments, no data to show which treatment was best, and no understanding that patients of similar age, heritage, and genetics might respond to certain drugs in the same way,” he says. “These two sisters had the same disease but very different medical trajectories.”
In the dawning era of personalized medicine, times are different. The availability of genomic information and the increasing use of electronic medical records (EMRs), combined with new methods of machine learning that allow researchers to process large amounts of data, are speeding efforts to understand genetic differences within diseases—including diabetes—and to develop treatments for them.
Bertsimas
Continue
reading