Health IT Analytics: ‘More Time Spent Shoveling Coal Than Steering the Ship’ | Healthcare Informatics Magazine | Health IT | Information Technology Skip to content Skip to navigation

Health IT Analytics: ‘More Time Spent Shoveling Coal Than Steering the Ship’

August 10, 2017
by David Raths
| Reprints
At Philadelphia Health IT Summit, population health execs describe challenges involved in feeding care management platforms
Click To View Gallery

As health systems and accountable care organizations establish the infrastructure to provide actionable data to care managers and front-line clinicians, they continue to struggle with messy data, incompatible systems and a lack of automation.

“We spend more time shoveling coal than steering the ship,” said Robert Kagarise, director of population health informatics and IT for the Delaware Valley Accountable Care Organization, the largest ACO in the Philadelphia area. Speaking during an Aug. 10 panel session on analytics at the Health IT Summit in Philadelphia, he said: “Feeding the beast takes more time than I would like. We want to shift our energy to looking at the data and navigating where we are going.”

Terri Steinberg, M.D., chief health information officer and vice president of population health informatics for Christiana Care Health in Delaware, agreed with Kagarise that supporting the plumbing of a population health effort takes considerable IT resources.

“None of the population health vendors can support the automation we need,” she said. Christiana has created its own Carelink CareNow platform for care managers, which enhances care coordination and includes a predictive analytics risk score for patients. Among the data sources feeding the platform are ADT feeds from the Delaware Health Information Network, the statewide HIE. The care manager may get a notice and contact the emergency department about a patient when the patient is still in the waiting room.

She said ideally the administrative processes feeding the care management platform would be entirely automated.  “We have developed our way out of that, but it means we have a lot of operations to support,” Steinberg said. “Our platform is expensive and we have to support it with volume.”

Kagarise said the Holy Grail is really to pull data from transactional systems and analytics platforms into one place that fits with the work flow of clinicians. “In three to five years, larger EMRs may have that capability, but they are not there today,” he said.

Kagarise, who used to work at Trinity Health, recalled efforts that failed because the analytics were not integrated closely with the work flow. “I remember that at Trinity we created 20 different predictive analytics and the chief medical officer was ecstatic, but two months later no one was using it because it was not in the EMR. That has not changed in the last five years. Workflow is the biggest problem. Ultimately we have to get everything under one roof with no issues around integration.”

Steinberg recalled a similar experience when Cristiana got tons of claims data to analyze and new reports to slice and dice. “I was so excited. I was like a kid in a candy store,” she said. All the doctors were going to log in and see how they were doing compared to their peers on different measures, she thought. Instead, “absolutely nothing happened,” she said. “Nobody had time to play with data. That was an important lesson: You need to tube-feed the providers and give them what you need them to digest in visualized reports.”

On the same panel, Robert Neff, director of innovative technology at Thomas Jefferson University and Jefferson Health System, said his group is focused on using data to improve the consumer experience at Jefferson. “Lots of hospitals have innovation groups. We are focused on how we improve consumer experience using digital technology.” He said that although data warehouses and data lakes are critical, Jefferson tends to look at its data with specific goals in mind, and often the data they need is not in the EHR or data warehouse; it might come from facilities or purchasing.

Another focus is real-time data analytics. EHRs are good at dumping out reporting data, he said, but often analytics executives are trying to change provider behavior with data that is months old. He said real-time data feeds have a more immediate impact. “When our providers treat a patient or enter a diagnosis code, they get feedback instantly,” he said. Emergency departments have dashboards with 55-inch TV screens. “Getting data out of the EHR in real time is a big challenge, but we are doing it.”

Steinberg said Christiana definitely uses “big data,” including socioeconomic data, to develop risk scores. “Once you know who is at risk, that is where little data comes in,” she said. “You need the boots on the ground. Large data sets are good, but it comes down to messaging the provider or patient. You can’t attend to 200,000 records every day. You need big data to drive you to the right place. Then you can use care management to drive provider behavior.”

She said value-based care is causing health systems to expose primary-care providers to analytics so that they can understand the characteristics of their panel and what is driving cost. For instance, they can look at similar procedures with large variations in cost. Orthopaedist A and B might have very different costs for knee replacement. The PCP is going to tell the surgeon: ‘I am at risk for this population and it impacts my bonus. I may not refer to you if you don’t change patterns because the outcomes are the same.’

As moderator Katherine Schneider, CEO of the Delaware Valley Accountable Care Organization, reminded the audience: “We have made some progress, but we are very early in this journey. We are just scratching the surface. De-fragmenting the data is one key to moving forward.”





Get the latest information on Health IT and attend other valuable sessions at this two-day Summit providing healthcare leaders with educational content, insightful debate and dialogue on the future of healthcare and technology.

Learn More