As healthcare moves toward value-based care and ultimately an at-risk payment environment, population health management is emerging as the delivery model required for success. Born before health reform, population health management is fundamental to the new system's "triple aim" — to improve health outcomes, lower costs and enhance the patient's care experience.
Population health management strives to achieve the triple aim "one patient at a time" by ensuring that people receive appropriate and timely preventive and chronic care, assistance navigating the system and resources to help them become more informed and engaged in caring for themselves.
It's a tall order. According to U.S. census data and the Centers for Disease Control and Prevention, the average primary care physician's panel of 2,500 patients includes 207 people with diabetes1 and 162 people with heart disease.2 Among patients over 20 years of age, nearly 25 percent smoke, 23 percent have uncontrolled high LDL cholesterol and 12 percent have uncontrolled hypertension.3 And, one in four Americans have more than one chronic condition. As a group, these individuals consume approximately 66 percent of total healthcare spending.4
Most people with multiple chronic conditions require extensive support to modify their risk factors and navigate the health system when services outside of the primary care physician's office are needed. Yet it is unrealistic for office staff and care managers to maintain personal contact with the thousands of patients who have "care gaps" and require care coordination and monitoring in a medium-sized practice.
To help scale their staff while pursuing a population health management strategy, healthcare organizations increasingly are turning to technology and automation to produce the best outcomes at an affordable cost.
How can you find clarity amid all the noise?
If technology and automation are prerequisites for population health management success, then how can a practice determine which technologies are best? These days, nearly every healthcare technology company claims to have a population health management solution. Logic and history dictate that not all of them will work.
I would suggest there are four key strategies that are critical to a successful population health management program. Each of these strategies can be translated into requirements for a technology platform.
- Fish smart; don’t boil the ocean
- Produce "actionable" data, not reports
- Make sure actions fit within caregivers' workflows
- Enable scale
Fish smart; don't boil the ocean
The current fashion in healthcare IT would have large health systems and physician groups deploy an enterprise data warehouse before pursuing a full-fledged population health management strategy. Data warehouses pool relevant data from electronic health records, enterprise resource planning, financial and other applications to produce a single "source of truth." The problem is they can take years to fully implement; yet health reform waits for no one.
Rather than boil the ocean with a data warehouse, organizations can use the databases they currently have to feed the analytics required for population health management. What are those analytics? To begin with, practices must be able to identify, visualize and attribute the population for which they are responsible. They must identify gaps in care across their entire population. They should be able to see the profile of the population by disease category, by admission patterns, ED utilization patterns and unnecessary care patterns. And they should be able to visualize all of this data at the level of the enterprise, facility and physician, as well as by payer contracts.
Accomplishing this with an EHR alone is difficult if not impossible. Many organizations are turning to specialized applications that complement the EHR and practice management systems to create a population-wide electronic registry. The purpose of such a registry is to assemble the latest data on the problems, medications and lab results of a patient population, including what services have been provided to each patient, where they were provided and the dates of service. By applying a set of evidence-based clinical protocols to the registry, an organization can determine which patients are overdue for particular types of preventive and chronic care, such as mammograms or diabetic eye exams.
Once a registry is in place, risk stratification software can classify a population by their level of health risk. That will show whether patients are high risk and need the personal assistance of a care manager; have less serious chronic conditions but require online support and other interventions, such as nutrition classes or group visits, to make sure they stay on track with their care plans; or are fairly healthy and simply need to be reminded about preventive care and good health behavior.
Data visualization is critical to make all of this data meaningful for caregivers. The most advanced analytic programs allow users to see how particular subpopulations are faring. For example, a scatter plot can reveal on one screen every diabetic or heart failure patient across multiple care providers and indicate which patients have their A1c levels under control. As one care manager in a Georgia practice recently put it: "I can look at all our diabetics across our 50 primary care providers and see where their A1cs fall. On just one screen, I can see thousands of patients and see who the outliers are."
Produce "actionable" data, not reports
While seeing is believing, a graph or a list alone is not enough to make lasting changes across a patient population. Healthcare is rife with sophisticated analytics tools that can extract, stratify and visualize a population to the nth degree, revealing new insights. In health systems and large practices, IT analysts can produce reports on virtually any topic a provider can dream up. But reports alone can't help a practice take the action required to engage patients and succeed at population health mangement.
What are the actions required for PHM success? The secret sauce is simple to communicate but hard to practice: care teams empowered to engage patients, one at a time and across cohorts.
To improve patients' overall health and reduce their risk of becoming acutely ill, primary care practices, for example, need to engage not only those patients who are out of compliance with recommended care but also those who are trending that way, such as patients showing signs of pre-diabetic metabolic disorders. They also must know about, track and engage patients for whom they are responsible – under contract – but who are receiving care elsewhere inside or outside their system.
Once patients are engaged, organizations then need to coordinate and manage their care and deliver targeted and personalized communications, alerts and interventions. Moreover, they need to be able to track and measure their progress against both internal and external performance metrics. Finally, they need to constantly update their performance scorecard with fresh clinical data, not dated claims data.
That is a lot of action to manage and maintain across an entire population.
If organizations lack a system that performs or supports these actions automatically, many patients fall through the cracks and may become seriously ill. Physicians and care managers can end up waiting for patients to remember to come in for care, which often doesn't happen. By contrast, when automation software detects patients who need services, it can automatically send out phone, email or text messages urging them to make appointments with their providers. Similarly, automation software can send timely alerts to assigned physicians and care team members to reach out to their patients proactively.
Make sure actions fit within caregivers' workflows
Just as healthcare is overburdened by shiny analytics tools that fail to enable meaningful action, nurses and care managers are often burdened with new applications that add complexity, not efficiency. Technology that doesn’t fit within an end-user’s existing workflow is bound to fail.
The best population health management automation tools are designed to make the care managers' everyday jobs easier, not harder. For example, population health managementsoftware can replace the spreadsheets and other manual tools many care managers fill with data from the EHR to keep track of high-risk patients. Not only can the software identify patients with care gaps, but it can also prioritize patients to help care managers quickly spot trouble. Based on their care gaps and health status, the software can highlight patients who have poorly controlled conditions with red icons, while those who are in less serious shape have green or yellow icons next to their names. Patients in trouble can then automatically be pushed to the top of a care manager's queue, saving valuable time and keeping patients from falling through cracks in the process.
Population health management applications should also add efficiency to other staff members. Front office staff and care coordinators, for instance, should be able to look at the provider schedule and, immediately under each patient’s name, see opportunities for their care to be improved. The software should immediately reveal if the patient is past due for a cholesterol test or mammogram or needs a colonoscopy.
Yet even a solution that lets caregivers take concrete action and fits easily within their workflows isn't enough to solve the PHM puzzle. The key word in the PHM approach is "population." Any technology that hopes to make the process easier must also make it scalable across an entire population — and more.
Make it scale
Getting technology to "scale" is an over-used term. To explain what it means for population health management technology, consider a typical primary care practice in which each physician sees 30 patients a day. With that patient load, physicians and their staff already have little time to make follow-up calls to these patients or to monitor the status of people who have been hospitalized or referred to specialists. Yet population health management requires that they not only do all of this, but also identify and communicate with patients who have not visited in a long time, engage people in maintaining or improving their health, and help educate those with chronic diseases on how to participate in their own care.
How do you turn one care manager into an army? Limited staff means organizations must be equipped to launch hundreds of health improvement programs simultaneously, each one aimed at a different subpopulation. For example, if the target cohort is 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.
But a scalable solution for population health management means more than expanding the capabilities of existing staff. As value-based contracting becomes more commonplace, organizations will need a population health management solution capable of scaling horizontally, across multiple commercial and government payers. As each payer launches unique quality-improvement initiatives, they require organizations to track and measure differing quality measures. Organizations have to align with multiple payers’ quality definitions and denominators — while performing analytics and predictive modeling across multiple clinical conditions.
Conclusion
Population health management completely changes a provider organization's business model and its approach to care delivery. To promote health and to help patients who are likely to become ill in the near future, an organization must be able to scale its patient engagement efforts to address the needs of every person in the population. The technologies best suited to help them accomplish that goal will meet the four requirements laid out here.
But even as an increasing amount of care delivery becomes automated, provider organizations should never forget the central importance of the physician-patient relationship. What makes automated messages or online educational materials effective is the simple fact that they come from the patient's physician. Physician-led quality improvement initiatives, enabled by data analytics and scaled up with automation, are what will move the needle on population health.
Karen Handmaker, MPH, is vice president of population health strategies for Phytel, a Dallas-based provider of population health management technology.
1 CDC, "National Diabetes Fact Sheet, 2011," accessed at http://www.cdc.gov/diabetes/pubs/pdf/ndfs_2011.pdf.
2 CDC, "Prevalence of Heart Disease—United States, 2005." MMWR. 2007;56:113-118.
3 CDC, "Prevalence of Uncontrolled Risk Factors for Cardiovascular Disease, 1999-2010," NCHS Data Brief No. 103, Aug. 2012, accessed at http://www.cdc.gov/nchs/data/databriefs/db103.htm.
4 CMS, accessed at http://www.cms.gov/Research-Statistics-Data-and-Systems/Statistics-Trends-and-Reports/Chronic-Conditions/Downloads/2012Chartbook.pdf.