Maine recently implemented a data analytics platform on top of its state health information exchange, enabling Maine HIE participants to predict readmissions and ER visits.
Ten years after kicking off, the Maine health information exchange (HIE) remains strong. Earlier this year, the HIE added a data analytics layer to its offerings, and recent statistics show participating hospitals are seeing returns on the effort. The analytics layer helps healthcare organizations in the state predict admissions, readmissions, ER visits and even a patient’s risk of heart attack, stroke, diabetes and death in near real time.
While many healthcare organizations use data analytics in some way, the interesting aspect in Maine’s case is that the data analytics platform — a contracted service — is being applied on top of the HIE, meaning healthcare organizations participating in the Maine HIE are able to work off the same data, said Cynthia Burghard, research director at IDC Health Insights, who is responsible for research on accountable care organizations.
The Maine HIE contains clinical data for nearly all of Maine’s 1.3 million residents, and new clinical data is uploaded nightly. In addition, care managers receive alerts almost immediately of any news surrounding the patients in their care — for example, if a patient is admitted into the ER.
“If I walk into an emergency room, anyone who’s associated themselves with me gets a message … within a minute of that event happening,” said Devore Culver, executive director and CEO of HealthInfoNet, at the Health IT Summit in Boston put on by the Institute for Health Technology Transformation. HealthInfoNet is the non-profit creator and operator of Maine’s HIE.
The number of alerts sent is impressive. For example, in May, patient care updates prompted the HIE to send 14,000 messages to caregivers, Culver said.
Predictive analytics part of new offering
The analytical platform accessing patient data in the HIE comes from HBI Solutions, a healthcare business intelligence company in Palo Alto, Calif. Using an algorithm, this analytical tool enables hospitals, ambulatory services and primary care practices to predict events such as admissions, readmissions, heart attacks and so on.
One of the first hospitals in Maine to pilot this analytics platform, and the first to sign a contract to use the approach full-throttle in March 2014, is St. Joseph Hospital in Bangor.
Before using the HBI solution, St. Joseph used BOOST, a guide many hospitals use to predict readmission rates. But once St. Joseph implemented HBI on top of the data in the HIE, “clearly [the new offering] was much better than BOOST,” said William Wood, M.D. and vice president of medical affairs at St. Joseph Hospital. He added that BOOST was time intensive and would often take a nurse about 15 minutes to go through a patient’s chart, answer questions and fill out forms. On top of that, BOOST used historical data, while the HBI tool is based off real-time data uploaded to the HIE every night.
When St. Joseph began using HBI’s analytics platform, not only did its readmissions fall below the state benchmark within a 30-day period, so did its admissions and ED visits, Wood said.
For example, in February 2015, the state’s benchmark called for healthcare organizations to aim for about a 23% ED return rate; St. Joseph achieved about a 19% ED return rate.
“That’s really remarkable,” Culver said. The HBI tool also allows St. Joseph to predict which patients will be the most expensive and then focus clinical efforts on those people, Wood said. In doing so, he explained, the tool saves the hospital money.
“Every time one of those individuals walks into the emergency room, it costs [the hospital] money,” Culver said. With predictive analytics, the hospital tries to, “redirect that person into the primary care structure and then bring forward the risks to help that primary care structure manage the patient and get the patient out.”
Culver added that this approach also allows the hospital to predict how much treating the patient population is going to cost the following year.
Sharing patient data key to Maine HIE success
All healthcare organizations participating in the Maine HIE can access the data in the exchange. The clinical information is collected from 31 of Maine’s 36 acute care hospitals and 326 ambulatory locations, including primary and specialty care practices and long-term care facilities.
Work on the Maine HIE began in 2005 when providers in the state agreed to not compete on the exchange of patient data, as well as to consolidate and standardize the data put into the health information exchange.
Without consolidation and without the concept of not competing for patient data, the HIE never would have worked, Culver said. However, “the decision to standardize data where possible leads us to where we are today.”
St. Joseph is also working on using the data analytics platform for other purposes, too, including monitoring real-time volume and market share data, creating key performance indicators and performing variation management — managing variations in outcomes, productivity, performance, patient flow and so on.
Before the analytics platform, Wood said St. Joseph didn’t have a tool the staff would use on a daily basis to identify patients who were developing risk or who were already at risk. But once the hospital began using the analytical platform, “[it] allowed us, in real time, to be making better decisions about that transition of care from the hospital to home.”
Though still in the early stages, Burghard said that analytics platforms on top of HIEs is “a growing trend where all of the HIE vendors and/or states that have put up HIEs need a second act — and analytics is the perfect solution for that because they’ve already done the hard work of pulling the data together.”