MDNews - Central New York

processing to transform dictation into structured data. And on a larger level, the Health Information Technology Standards Committee of the Office of the National Coordinator for Health Information Technology (ONC) is involved in developing a standardized, interoperable method for transforming documents into data. "You look at how the EHR is storing information," says Charles Kennedy, MD, Chief Executive Officer, Accountable Care Solutions from Aetna, who serves on the committee. "[We are] developing standards so systems can consume human-readable data and turn it into machine-readable data." Data into Intelligence: the Alchemy of Analytics Medical groups need to keep patients healthy and also make a profit. Transformed into business and clinical intelligence, data promises to help them do just that. "Organizations require sophisticated analytic tools so they can … improve the quality of care and reduce avoidable hospital admissions [while also applying] financial analytics to measure the cost of care delivery and use … healthcare resources wisely," says Ken Terry, writing for the Institute for Health Technology Transformation (iHT2). So, where do these tools come from? Very small practices just meeting the minimum ACO requirements may rely on manual searches and hand calculations, perhaps combining them with the compassion and excellent memory of a longtime administrative assistant to remember "which patients need care when." However, such a model is not scalable. Some EHR vendors supply analytics engines, such as Epic's Cogito. Many of these engines offer a palette of tools, including patient registries, decision support and dashboards. Larger health organizations may build their own analytics systems. In other cases, health organizations that align with insurers to form ACOs may draw upon the insurer's preexisting tools, such as Aetna-affiliated ActiveHealth Management, for identifying risk and managing health across large numbers of patients. Population Health Management Once an ACO has turned data into a broad understanding of the health of its member patients, what next? How does a practice or health system turn that understanding into improved outcomes? Some analysts advocate for ACOs to use predictive modeling to identify which patients will suffer a fall, be admitted to the emergency department, and so forth. Those patients are then "headed off at the pass" with interventions, such as a visit from a social worker or closer health monitoring. According to iHT2, "high-risk patients generate the majority of health costs [and] only 30% of high-risk patients today were in that category a year ago." Therefore, prediction matters to both the clinical and business bottom lines. For those who favor the predictive modeling approach, the relationship between analytics and patient messaging becomes obvious. As the American Medical Group Association notes, ACOs bear the risk for all their patients, even those who "are not cooperative or engaged" in their care. There is a "reminder" or even "inside marketing" aspect, in which patients with predictable health issues are selected for additional persuasion to engage in healthy behavior, as well as extra support, such as testing, counseling and the like. In fact, under a fully optimized ACO model, marketing may expand into promoting healthy behaviors and primary care use by existing patients, as well as driving specific service lines. Not a ll ACO experts ta ke this approach. Behavior in the aggregate is possible to predict, but predictive analysis doesn't always translate well to individuals. R ather tha n forecasting which members of a risk categor y w ill have certa in outcomes and responding with interventions, Dr. Kennedy suppor ts a more streamlined, algorithm-based approach to guiding health outcomes among large numbers of people. "On a population basis, you can predict how people will behave with good accuracy; the greater challenge is to predict what happens to individual patients," Kennedy says. "It would be good to know which CHF [congestive heart failure] patients will be readmitted; you'd want to apply your resources to those less likely to thrive. We think the better use of technology isn't so much predicting what's going to happen but making sure when you know the patient has the disease that the interventions proven by research actually happen." Kennedy recommends an algorithmic system based on sound research and best practices to budge outcomes in the aggregate, rather than investing analytical resources on a patient-by-patient basis. This moves the field of agency back toward providers, allowing broader influence over outcomes, and skirts some of the difficulties involved in predicting — and subsequently influencing — behavior at the individual level. "Accountable care offers physicians, especially primary care physicians, an opportunity to take back the practice of medicine," Dr. Kennedy says. "I do think physicians will find the eventual outcome more satisfying, better for patients and better for themselves financially." W hat role does analytics play in your ACO? Join the di sc u ssion at www.MDNews.com. ■ MDNEwS.COM ■ MD News Central New York |9

• Cover