It's not unusual for a health care provider to mine the medical records of its patients for research. But gaining access to the medical records of virtually an entire county is a far greater challenge.
Since 1966, the Mayo Clinic has collected as many medical records as possible in Minnesota's Olmsted County to generate powerful studies that help save lives.
The Rochester Epidemiology Project has a treasure trove of close to 600,000 medical records. The health data have generated more than 2,000 studies and provided groundbreaking information on a variety of conditions, including heart disease and attention deficit hyperactivity disorder (ADHD).
After nearly 50 years of success, the project is expanding to include patients from seven other southeastern Minnesota counties, which will help researchers broaden its database.
Having access to a large and growing pool of medical records gives researchers many more study options, said Jennifer St. Sauver, scientific manager of the project.
"The great thing about the Rochester Epidemiology Project is that it's not confined just to one particular disease area or one particular condition," she said. "Instead, since it's just capturing all of the health care received by this local population, you can do studies of virtually any condition you can think of."
Over the years many of the project's findings have attracted national and global interest. For example, Mayo Clinic nicotine researchers scanning the database found that new cases of heart attacks declined by 33 percent in Olmsted County after smoke-free workplace laws were implemented in 2002.
Another well-known study revealed that children who had multiple exposures to anesthesia before age 3 had an increased risk of ADHD in research that is ongoing.
"They're following up that study now with a five-year grant to try and focus on understanding whether the anesthesia itself is a strong risk factor for later behavioral problems, or whether it's the type of anesthesia or the dose or whether it's other conditions that these children might have," St. Sauver said.
But not all studies that use the database are as powerful as they could be. St. Sauver said that's why researchers need medical records from more people.
"To do a study on, for example, pancreatic cancer in Olmsted County, just focusing on this population is hard to do because it's a rare cancer and it just doesn't affect enough people," she said.
Mayo Clinic researchers are starting to add medical records from affiliates in Dodge, Fillmore, Goodhue, Houston, Mower, Wabasha and Winona counties.
The hospital and clinic system also has made contact with other regional providers outside of its health system to see if they are interested in participating.
Dentists, optometrists and chiropractors have been encouraged to join the project, too. Bruce Trulson, a dentist who runs the Trulson Dental Clinic in Stewartville, Minn., said his practice has been a member of the Rochester Epidemiology Project for two years.
"You know we're part of the body, and there is more and more evidence that there's some close links between dental health and our medical health," he said.
Heart disease, diabetes and pregnancy, he said, can be strongly influenced by dental problems. "There's some pretty compelling research that shows that if a pregnant mother has chronic periodontal disease she's at a much higher risk of giving birth to a low-weight baby," Trulson said.
More than 80 percent of Trulson's patients have agreed to add their dental records to the database.
But high participation isn't the only goal as the project expands.
Mayo Clinic researchers also aim to improve racial diversity within the records pool. That would address a longstanding criticism of the database, as Olmsted County is predominantly white.
St. Sauver acknowledges the challenge. "We'll probably never look like south Texas because we are in Minnesota and we are in the upper Midwest. But, you know, south Texas will never look just like Olmsted County or the rest of Minnesota," she said. "So that doesn't mean that studies done here can't be generalizable to other areas."
In time, the demographics represented in the database will likely change, St. Sauver said.
She points to the Rochester public schools, where minority children now make up nearly 30 percent of the elementary school population. Within 30 to 40 years, she said, those children and their offspring likely will represent a much more significant portion of the people in the Rochester Epidemiology Project.