How Big Data Helped Cut Emergency Room Visits by 10 Percent

When a man is rolled into an emergency room with acute abdominal pain, a doctor he’s never met must quickly try to figure out the cause of the problem: Kidney stone? Appendicitis? Triple cheeseburger? If the physician had a way of knowing that the man complains of aches and pains all the time, that a recent CT scan revealed no problems, and that a few days earlier a doctor at another ER had sent him home with an antacid and instructions to follow-up with his family doctor, the examination would take a lot less time and likely cost a lot less money. “All of a sudden, it changes the whole nature of the conversation with the patient,” says Dan Lessler, the chief medical officer for Washington state’s Medicaid program.

An experiment under way in the state is starting to show how giving doctors that kind of information can reduce the number of people who repeatedly show up in emergency rooms with minor ailments such as upset stomachs and headaches. Many are low-income patients covered by state Medicaid programs—which means taxpayers pick up the bill. Almost all ERs “have a few patients who have the potential to abuse the system,” says Mark Reiter, president of the American Academy of Emergency Medicine. “I had a patient come in over 300 times one year.” In Washington, patients who went to the ER more than four times a year made up a fifth of all ER visits paid for by Medicaid. So in June 2012 the state began requiring all ERs to put details about patients they treat into a statewide database, allowing physicians to track patients’ visits even if they frequent more than one hospital.