Balance, workflow hurdles remain for provider use of patient-generated data

Despite advances in health information technology, personalized medicine remains a challenge for many providers. From finding the right balance for using incoming data to overcoming workflow issues, no magic bullet exists to ease the hurdles of hyper focusing care efforts.

During a panel discussion at Health Datapalooza in the District of Columbia this week, clinicians talked about some of those challenges, as well as some of the ways they're succeeding in delivering more personalized care.

Stanford Clinical Assistant Professor of Pediatrics Rajiv Kumar (pictured right), for instance, talked about the need to integrate patient-generated information into electronic health record systems, so as not to impede provider workflow. He said his organization is using Apple's HealthKit, which enables clinicians to view such information in their EHR; it also allows them to manipulate data and build custom alerts.

"The biggest obstacle to getting your physician or provider to look at patient-generated health data is the physician," Kumar said. "Patients, for the most part, want this, or they're not as scared of it." For providers, he added, there's a fear of possibly missing a new piece of information and then being held accountable for that.

Kumar, who specializes in treating patients with diabetes, called that disease, in particular, very overwhelming in terms of data.

"It's a lot of self-management," Kumar said. "Once patients are diagnosed, we admit them to the hospital where they receive intensive education. We teach them how to check their blood sugar, which they have to do at least four times a day, forever. They have to learn how to count all of the carbohydrates in every meal that they're going to eat, forever."

To that end, he called looking at discrete points in time for care worthless, and said that being able to continuously monitor patients via Apple's tools is more effective. However, he worried about the reimbursement ramifications of losing out on traditional three-month check-in visits, saying that assessing patient-generated data every two weeks could render such visits obsolete.

Still, Seth Bosker (pictured left), an associate clinical professor of pediatrics at the University of California, San Francisco School of Medicine, said patients are much bigger than just their data. He told a story of a mother of a patient who shared information from an app with him that pinpointed each day over the course of a month when her child suffered from a seizure.

"This mother was showing me this app, not just to share her data; she wanted to share her experience," Bosker said. "She wanted to share with me what her month had been like. … To get to truly personalized care, we need to connect on more than just bits and bytes."

Bosker added that any data upon which clinicians are making medical decisions needs to be in the medical record, but said his EHR doesn't have the infrastructure to handle real-time incidents based on patient-generated data. For that reason, he said, UCSF does not import a lot of that data into its EHR.

"We have some primitive clinical decision support tools in our EHR that allow us to set thresholds based on one number, but we don't have the algorithms yet developed to do that intelligently with high sensitivity and specificity," Bosker said.