Using novel methods to extract data from EHRs, researchers at two Boston hospitals identified three factors that potentially may predict PCI readmissions. They reported their findings Nov. 18 at the American Heart Association scientific session in Chicago.

The 30-day readmission rate for PCIs reached nearly 12 percent in 2011, according to an analysis of the National Cardiovascular Data Registry’s CathPCI registry. These readmissions add costs for Medicare and private insurers and burden patients. Medicare alone shells out $360 million a year on readmissions, Jason H. Wasfy, MD, of Massachusetts General Hospital in Boston estimated in an interview with Cardiovascular Business.

“Thinking about how to reduce PCI readmission offers an opportunity to both reduce healthcare costs and improve the quality of care we deliver,” he said.

He noted that while risk models based on registry data exist for PCI readmission, identifying vulnerable patients remains a challenge. “We wanted to understand better how to predict prospectively who is going to come back to the hospital and not come back to the hospital," Wasfy said.

Wasfy and colleagues took advantage of the robust EHR and high-volume PCI centers at Massachusetts General and Brigham and Women’s hospitals to perform an analysis of nearly 9,300 PCIs. They identified potential risk factors through literature searches and physician insights and used semi-automated techniques to extract unstructured and semi-structured data from the records.

In an analysis that matched 888 readmitted patients and 1,776 patients who were not readmitted, they determined that anticoagulation, the number of previous emergency room visits and anxiety were independently associated with readmission. Anxiety may offer physicians a golden opportunity to reduce readmission rates, Wasfy said, because it is modifiable and actionable.

Interventional cardiologists excel technically but there may be room for improvement in patient education, said Wasfy, who is the co-chair of the Acute Myocardial Infarction Clinical Care Redesign Committee at Massachusetts General’s Institute for Heart, Vascular and Stroke Care. “To do a good PCI is not to just to do the procedure right,” he said. “We need to educate patients about what it means to have heart artery disease and make sure they have the resources to manage their conditions and good ways to access their physicians to ask questions.”

Wasfy and colleagues previously have demonstrated that more than 40 percent of PCI readmissions may be preventable. Their next steps include testing the new variables in a risk model prospectively to evaluate whether they improve the model’s predictive ability.