Physician-led quality improvement initiatives, enabled by data analytics and scaled up with automation, will move the needle on today's population health efforts.

With value-based reimbursement just around the corner, more and more healthcare providers are moving toward a patient-centered medical home delivery model. To the extent that they share savings or financial risk with payers and are required to meet quality goals, these organizations must learn how to operationalize population health management (PHM). In other words, they must try to keep their patient populations as healthy as possible while reducing the need for costly medical interventions.

PHM requires providers to engage all of their patients continuously, not just when they come into the office for treatment. They must track patients' health status, identify care gaps, engage patients so they receive needed preventive and chronic care, provide care management between visits and persuade patients to monitor their health risk factors and participate in managing their own care.

PHM is challenging without automation. Healthcare organizations must look beyond manual efforts, including the hiring of care managers, to care for an entire population on a one-to-one basis. Some providers have launched PHM pilots, hiring a few care coordinators to conduct manual outreach and patient coaching. While many have had good results, they focus on only a fraction of their patients outside of scheduled visits unless they have automation tools that enable them to scale the approach to the entire population. They must also leverage information technology to aggregate and analyze patient data and to design interventions for population-wide health improvement.

Clean, actionable data
Data is like oxygen to PHM initiatives. You need as much data as possible on your patient population, in a reliable and usable state. Regardless of how brilliantly designed your automation tools are, they won't work as intended unless your data management capabilities are rock solid.

To execute on PHM effectively, providers must aggregate data from multiple sources in a patient registry designed to keep track what has been done for individual patients and what services they need. When analytic engines are applied to registry data, they can provide breakdowns of various subpopulations, based on demographics, health conditions, medications, lab values or other factors. Analytics can also identify care gaps and stratify a population by health risk.

To make this data actionable, the system must send providers and care teams relevant alerts via their EHRs. Care managers must also have the automation tools they need to intervene with patients cost effectively. These applications should be part of a patient engagement strategy that proactively seeks to prevent people with incipient health problems from becoming sick enough to need more costly care.

This is a dynamic challenge that requires organizations to keep track of patients' changing health status. At any given time, about 3 percent of patients require hands-on care management, but 70 percent of the individuals who will need that level of care a year later will be different people. So provider organizations must pay careful attention to what's going on in their population beneath the water line, because it will impact patient outcomes as well as the future cost of care.

To fully address all patients' needs, organizations must be able to launch hundreds of health improvement programs simultaneously, each one aimed at a different subpopulation. For example, if you select as your target cohort all diabetic patients with an HbA1c >9 and a BMI >35, a PHM automation suite that integrates analytics with care delivery should be able to configure a campaign to address that cohort instantly. Such a campaign might include sending those patients automated phone messages, emails or texts suggesting that they make an appointment with a diabetes nutrition counselor.

Communicating with patients on mobile devices is rapidly becoming critical to this outreach capability. While email is still important in non-visit care, many providers and patients no longer want to sit down at a PC and log on to exchange information. Texting is particularly useful for reaching younger patients. Nevertheless, because some patients use different technologies and have different communication preferences, messaging must be multi-modal.

Finally, provider organizations should never forget the central importance of the physician-patient relationship. Even as an increasing amount of care delivery becomes automated, the fact that automated messages or online educational materials come from the patient's physician is what often persuades that patient to listen to the messages or to interact with the content. Physician-led quality improvement initiatives, enabled by data analytics and scaled up with automation, are what will move the needle on population health.

About the author
Steve Schelhammer is chief executive officer at Phytel, a leader in automated, physician-led population health improvement. To learn more about Phytel, go to