News

Pop-up blocker in the examining room

Clinical reminders become more helpful:

New information filter model for decision support

A patient and a doctor walk into an examining room. The physician opens a computer, and on the screen is a whole dashboard with the patient’s medical history.

That’s one advantage of using an electronic medical record like the Veterans Health Administration and other health systems do.

Clinical reminders pop up on the screen too, alerting the clinician of upcoming or overdue medical labs, procedures, or exams, or possible drug-drug interactions. They may alert the clinician that the patient needs to be screened for depression, or tobacco use. The patient may have had high blood pressure last time, or a diabetic may be due for an eye or foot exam.

The Veterans Health Administration expects its clinicians to resolve at least 85% of clinical reminders – meaning acknowledging the reminder by making it go away from the screen.

The intent of reminders is to make sure nothing important falls through the cracks. It’s part of what is called a decision-support system, designed to help clinicians make good, thorough decisions about their patients’ care.

But clinical reminders can be unwieldy when they are irrelevant.

“We have developed some elegant ways to support clinicians’ decisions,” says Center Director and Investigator Brad Doebbeling, “but there are too many alerts and reminders, and they are not prioritized for the individual patient.”

Doebbeling says seat belt reminders can crop up when a physician is caring for a critically ill patient. Or up pops a mammogram alert when it’s a man sitting on the examining table.

“Every time I talk to clinicians,” Doebbeling says, “this is what they’re talking to me about.”

While clinical reminders can be helpful, they can also waste precious time. It would take six hours, research suggests, for the clinician to cover all the currently-recommended preventive bases.

A team of Center investigators is starting to write software that will help separate the wheat from the chaff.

Their information filter model will cull reminders that do not apply to a certain patient, while filtering in the most important individuated information. This model will not only decrease undesired, useless and irrelevant reminders, but also decrease the workload for healthcare providers while improving overall computerized clinical reminder information quality, and eventually improve the quality of care significantly.

The team’s primary scientists are Doebbeling, Center investigators Mindy Flanagan and Alan Zillich, and a group of Center affiliate investigators, all Industrial Engineers from Purdue University - Sandra Wu, Mark Lehto and Yuehwern Yih.

“We are taking an epidemiologic perspective,” says Doebbeling. “If we have the patient at the clinic for an appointment, how can we have the most impact on that person’s life in that one opportunity?”