Seizing opportunities in big data

BOSTON—Healthcare data are diverse, and range from claims to EMR data to genomic, lab, imaging and continuous physiological data. How to deploy big data analytics in healthcare and successfully move technology into practice is on the mind of Suchi Saria, assistant professor in computer science at Johns Hopkins University, who spoke at the Big Data Healthcare Analytics Forum on Nov. 21.

In 2013, Saria was part of a group of thought leaders brought together by the Gordon and Betty Moore Foundation to look at harnessing big data in healthcare using analytics to identify high-risk and high-cost patients.

From this, she shared near-term opportunities to seize as interoperability and technology continue its progress:

1. Hospital readmissions: With penalties in place, a focus on hospital readmissions makes financial sense. Interest is booming in the research world, Saria said, noting that in 2009, 56 research papers studied hospital readmissions while in 2014, this number jumped to 1,422 research papers.
2. Managing higher cost users: Five percent of patients account for 50 percent of resources, and most often these patients have co-morbidities. Already, many solutions are in place to identify such patients based on risk algorithms. Johns Hopkins’ Adjusted Clinical Groups (ACG) Case-Mix System, which is a large-scale population health system, can bring tailored interventions to case managers. The school also is working on an app that measures gait, tremor and other small day-to-day changes that can detect patient deterioration before major episodes take place.
3. Triage: This entails reducing overutilization by targeting resources based on need; smarter ways of transferring patients in and out of the ICU; and planning discharge. With analytics, now it is possible to predict which babies at three months of life will have major mortalities downstream, changing how patients are treated.
4. Adverse Events: This involves utilizing analytics to predict which patients are at risk of adverse events, such as renal failure or infection.
5. Decompensation: This helps healthcare workers identify patients on the verge of disease progression and intervene sooner. Risk scores developed at Beth Israel Deaconess Medical Center, for instance, can identify patients before they go into septic shock.
6. Complex Diseases: There is an opportunity to tailor interventions to unique individuals with complex diseases (such as rheumatoid arthritis, inflammatory bowel disease, etc.), which are the costliest to manage. Two to 4 percent of drugs are expected to make up 50 percent of the costs by 2018, she said.