Healthcare big data goes from big to great
Not only is data getting bigger, it's getting smarter.
For example, health insurer WellPoint is branching out from simply looking for gaps in coding, thanks to a combination of better data and more advanced algorithms. Now, the company can look for gaps in care as well, Patrick McIntyre, the company's senior vice president for healthcare economics, explained this past May at SAS Institute's 11th annual executive conference on health analytics.
About 2.5 million of WellPoint's 37 million enrollees have insurance tied to some sort of value-based reimbursement model, McIntyre said, and the Indianapolis-based payer shares reports with providers whenever there is risk-sharing. "We use analytics and reporting to create economies of scale for all of the provider communities we work with," he said.
WellPoint performs both retrospective analyses of claims – a more traditional form of data mining – as well as proactive analysis of care gaps. This helps the company coach providers on better coding and service delivery.
Health insurers, of course, historically have been met with mistrust and suspicion when they reach out to members and providers. With their vast data collections, that is changing. "There isn't a magic bullet, but it's really bound in trust," McIntyre said.
“We need to provide the right care to the right patient at the right time," he added. "I think analytics is going to be the differentiator."
Making sense of the unstructured
Mark Pitts, VP for enterprise informatics data and analytics at Highmark Health, the new, Pittsburgh-based parent company of Allegheny Health Network and health insurer Highmark, offered similar sentiments. As someone with years of experience at payers, a "primary challenge" for Pitts has been how to influence individual behavior to promote better health and save money.
Today, with the advent of "text analytics," organizations like Highmark can make sense of vast stores of unstructured data, not just information entered in a discrete format. (Pitts called this the "bag of words" method.)
According to Pitts, computers now can look for "term concurrence" across multiple documents to search out patterns, such as evidence of patient dissatisfaction, according to Pitts, so people don't have to flip through hundreds of pages in hopes of stumbling across something meaningful. "Have machines find things," he said.
Allegheny is getting ready to bring this technology to the provider side. For example, Pitts said, the length of a nurse's progress note often correlates with illness severity. By paying attention to patients with particularly detailed notes, the health system might be able to prevent medication errors, escalation of acuity and even hospital readmissions, he suggested.
Next wave for Kaiser data
Like Highmark, Kaiser Permanente is both healthcare system and health plan. That massive organization's EHR contains something in the neighborhood of 10 petabytes of data, according to Terhilda Garrido, VP for health IT transformation and analytics, making it ripe for big data technologies.
Garrido wants to gather "patient-reported outcomes" after each encounter, including what patients say on social media, data pulled from medical devices and patient satisfaction ratings. "That, for us, represents the next wave of the continuum," she said.
Physicians can order questionnaires through KP's My Health Manager patient portal. "It's a little clunky," Garrido said, but noted that it is the first stage of what probably will be a long effort. Later, there may be auto-collection of data from mobile and home-based devices, adding another way to track and measure outcomes.
Mostashari on gauging outcomes
If Kaiser is successful, former national health IT coordinator Farzad Mostashari, MD, would be pretty happy.
"Most of which determines our outcomes isn't what happens in our office," Mostashari, visiting fellow of the Engelberg Center for Health Care Reform at the Brookings Institution in Washington, said during the opening keynote of the SAS event.
"One of the things that drove me crazy in medicine is that I never got any feedback," Mostashari said. According to the former national coordinator, "99.999 percent of the time, we have no idea what we get" for all the money spent on healthcare.
SAS Chief Medical Officer Graham Hughes, MD, echoed some of these sentiments. "Ninety-nine percent of patient care takes place outside of traditional care settings," he said, emphasizing the importance of collecting and analyzing data from patients' everyday lives to close gaps in care and personalize treatments.
"Maybe we start to think of every disease as a rare disease?" Hughes wondered aloud as he discussed the potential of big data to help individuals make healthy lifestyle choices outside the sterile, controlled environment of a hospital or clinic.
"We're starting to get past the PowerPoint page and into some real situations," Hughes said.