Michael B. Matthews
Requirements for success with population health management include a high degree of clinical data integration across the healthcare organization's care enterprise and automated tools for engaging patients in their own care.
At a Glance
Online tools for automating population health management can help healthcare organizations meet their patients' needs both during and between encounters with the healthcare system. These tools can facilitate:
- The use of registries to track patients' health status and care gaps
- Outbound messaging to notify patients when they need care
- Care team management of more patients at different levels of risk
- Automation of workflows related to case management and transitions of care
- Online educational and mobile health interventions to engage patients in their care
- Analytics programs to identify opportunities for improvement
Impending changes in payment are prompting healthcare organizations to take a greater interest in population health management, a systematic approach to ensuring that all patients receive appropriate preventive, chronic, and transitional care. In essence, providers engaged in this new approach to health care enhance the cost-effectiveness of care delivery by focusing not only on meeting the needs of people who are sick or in immediate need of care, but also on ensuring the wellness of their entire patient populations.a
As part of this new approach, primary care practices organized as patient-centered medical homes assume greater responsibility for managing and coordinating the care of patients. As a result, population health management lowers hospital admissions and procedures. Hospitals that pursue this approach, therefore, must contend with the challenge of reduced revenues from inpatient care. This challenge will soon become unavoidable as the growing prevalence of value-based purchasing and risk contracting compels hospitals to make population health management a key component of their business strategies.
Population health management requires a healthcare provider not only to track the steps it has taken to identify and meet the healthcare needs of its patients, but also to monitor their health between visits or episodes of care so it can intervene proactively, give patients appropriate support, and engage them in their own care.
Clearly, population health management represents an inversion of today's business model, which is to generate more revenue by delivering more services. Population health management can generate increased revenue, but in a different way: It can boost the bottom line of physician practices-and the health systems that own them-by bringing patients back into the office for necessary preventive and chronic care. And to the extent that hospitals and physician groups enter gain-sharing or risk-sharing arrangements, it can help hospitals to increase revenue in the long run, as well.
To measure the success of this strategy, healthcare finance managers should look at the total margin of care across their enterprises. An important component of that margin is the cost of managing a patient population. If a healthcare organization hires too many care managers or spends too much on inefficient methods of identifying and communicating with high-risk patients, these excessive costs can have a negative effect on the system's bottom line. Fortunately, by using new health IT applications to automate many of the functions of population health management, organizations can avoid these costs-and achieve greater efficiency of care management, as well as more cost-effective patient monitoring and outreach.
Role of IT
Although population health management focuses partly on the high-risk patients who generate the majority of healthcare costs, it systematically addresses the preventive and chronic care needs of every patient. Because the distribution of health risks changes over time, the objective is to continuously identify the factors that make people sick or exacerbate their illnesses. The importance of this approach becomes clear when one considers that among the patients who are the most expensive to treat today, a high percentage may have incurred only relatively modest healthcare costs in the previous year.
Health IT and automated communication systems are foundational elements of population health management. Although human interaction is essential for managing high-risk patients, all routine aspects of population health management should be automated. For example, the tasks required to monitor a population's health, and to perform outreach and education, are too time-intensive to be accomplished efficiently using manual methods alone.
Documentation in electronic health records (EHRs) digitizes data to facilitate their use in population health management. Moreover, EHRs provide data elements that cannot be extracted from billing information. If disparate EHRs are linked together by a health information exchange (HIE), providers can use the EHRs collectively to track patient care across care settings. But many of today's EHRs lack key components that are necessary for population health management, including robust registries and outreach capabilities. Patient registries, which are lists of individual patients, the services provided to them, and the results of lab tests, are necessary to identify care gaps and analyze the care provided to particular population segments, such as diabetic patients with HbA1c values greater than 7.
This new approach also requires certain types of workflow and care management automation that are beyond the capabilities of an EHR. For example, most EHRs lack the ability to trigger messages to patients who are overdue for certain kinds of preventive or chronic care. And care alerts in EHRs are insufficient for care managers to identify all high-risk patients or those who might become high risk.
Other types of health IT applications, when used in conjunction with EHRs, can provide analytic capabilities to measure population health trends and identify areas for organizational improvement. They can also reduce the work and expand the capacity of care managers. And web-based educational materials and health risk assessment tools can engage patients while indicating their care gaps.
IT and Clinical Integration
Clinical data integration is required both for care coordination and for effective population health management. The reason is that patients receive care from multiple providers who are not necessarily in contact with one another. A physician in one practice or organization may be unaware that a physician in a different organization has prescribed a particular drug to a patient. If a patient receives a flu shot at a local retail pharmacy, his or her physician may not know about it. Providers that have broad access to such data can perform much more accurate population health tracking and analysis.
Interoperable information systems are the Holy Grail, yet many of today's EHRs can exchange little, if any, structured clinical data. Community HIEs can help fill the void, but exist in only a limited number of markets. Private HIEs are growing more rapidly, but they cannot provide data on care that patients receive outside of the enterprise.
Where HIEs have achieved a degree of maturity, capability, and provider engagement, they can be invaluable tools for supporting population health management. Not only can emergency department (ED) physicians use them to look up patients' histories when they present, but referring physicians and specialists can easily exchange notes through HIEs. Hospitals can use HIEs to send patient information to postacute care providers, although data can rarely flow to the same extent in the opposite direction, because postacute care facilities tend to lack comparable IT systems.
Automated Patient Tracking and Outreach
EHRs are designed to assist providers in caring for individual patients, not to manage a population's health. When a patient visits a physician and the provider opens that patient's record, health maintenance alerts pop up in the more advanced EHRs. But canned prompts for preventive and chronic care are fairly basic and customization is difficult. Most important, most EHRs do not prompt providers about care gaps when a patient is not in the office.
A growing number of physician groups are using the electronic registries described previously to track their patients' health care and alert patients when they need health maintenance or chronic care services. When these registries are interfaced with EHRs, they can also prompt providers about care gaps. The registry data come from several sources, including EHRs, practice management systems, and lab systems. Where EHRs are interfaced with HIEs, clinicians normally review the HIE data from outside sources before adding the data to their EHRs.
Registry prompts are based on nationally accepted clinical protocols that particular groups may refine to fit their own practice patterns. Organizations must apply the same security procedures they use for EHRs to these registries, since they contain individually identifiable patient data.
Some practices combine these registries with patient outreach programs. By applying clinical protocols to registry data, these programs can send automated phone, email, or text messages to patients, telling them to make an appointment with their physician.
Practices that do not have EHRs can use the same approach to patient outreach, based on registries populated by billing and scheduling data from their practice management systems. Clinical protocols can be applied to these administrative data. But an EHR supplies richer data that allow providers to target patients more precisely. For instance, a protocol for diabetes may say that patients should visit their provider every six months. But if a diabetic patient has an HbA1c greater than 9 or is obese, perhaps he or she should come in every three months. This information resides in EHRs, but not in billing data.
Automated Care Management
Although the goal of population health management is to deliver excellent care to every patient, the best way for providers to lower population health costs as much as possible is to pay special attention to patients who pose the highest risk of using a high percentage of healthcare resources (e.g., in a given year, it is possible for just 5 percent of a population to use as much as 50 percent of the resources needed to treat that population).
Manual approaches offer a limited solution. Many healthcare systems that are focused on managing population health are hiring substantial numbers of nurse care managers to manage high-risk patients.b Although this solution should be part of any approach to caring for the sickest patients, nurses are an expensive resource, and it is unlikely that they will be able to find time to contact any patients other than those who already are quite ill. As a result, those who are not yet so sick do not receive care that could prevent their conditions from worsening.
Automation can circumvent this problem effectively and cost-efficiently. It starts with "risk stratification," a technique that classifies patients according to their health status and other factors and their chances of becoming sicker in the near future.
When applied to registry data, risk stratification applications show which patients are high risk today and which ones will probably move into that category soon unless they receive proactive interventions, thereby enabling care managers to prioritize their case loads. Care managers can use registry-based reports to quickly see where each patient's care gaps are and determine which patients need extra assistance. Such population health data enable care managers to make the most effective use of their time. As a result, they can help more patients, so fewer care managers are required to achieve the same outcomes. From the organizational standpoint, that translates into an improved bottom line.
To assist in risk stratification, providers also can ask patients to complete online health risk assessments (HRAs) to identify their needs and show where they fit in the risk classification. The most effective HRAs also include questions to assess the patient's willingness to take steps to address the identified risks.
Based on the experience of organizations that are managing population health, about 3 percent of patients have a high need for support from clinical care managers.c Care managers not only should be able to receive daily reports on such patients assigned to them, but also should be provided with electronic tools to help track, communicate, and intervene with these patients. Such tools include home-based or mobile devices that allow for remote patient monitoring, thereby enabling care managers to track vital signs and other clinical data and respond to alerts on an ongoing basis.
Just below the highest-risk, sickest category are patients who have chronic conditions but are not yet very sick; these individuals can receive automated messaging that tells them when to visit their physicians for testing and treatment. They should also receive tailored educational materials, as discussed below. The same is true for patients who are fairly healthy, but need preventive care at regular intervals.
Automated Patient Engagement and Care Management
Patient engagement is crucial to improving population health because patients with chronic diseases-which generate 75 percent of healthcare costs-often manage their own conditions.d Equally important, healthy patients must learn how to care for themselves so they don't get sick and require costly health care. Research shows that engaged patients have better outcomes and produce measurable cost savings.e
Here again, automation is the key. Simple handouts of educational materials do not necessarily affect health behavior, especially with patients who have low health literacy and do not understand their physician's instructions on self-care or taking medications. Online patient education materials, in contrast, may be multimedia and interactive. They show patients what is required of them and encourage them to submit questions online. Such materials are available both for postsurgical care and for care of chronic conditions.
When combined with automated patient communications, these online educational materials can be a powerful tool to motivate patients. Physicians can put in a standing order for particular education pieces to be directed automatically to patients at various points in the care process.
Online HRAs also can engage patients by making them aware of their care gaps. Because getting patients to fill out long forms can be difficult, it is best to break HRAs into bite-sized chunks dealing with specific areas, such as smoking, obesity, and diabetes.
The new mobile health tools can also promote patient engagement. Mobile health apps can encourage people to become fitter, to lose weight and exercise, and to track blood sugar, blood pressure, weight, diet, and other indicators. The data generated by these devices can motivate patients to better manage their diabetes, heart failure, or other conditions. Moreover, transmission of such information to the patient's physician or care manager can alert care teams of the possible need for more, human intervention.
Automation of care management also can improve transitions of care after hospital discharge and prevent readmissions. Automated messaging to discharged patients can ask them whether they understand their instructions and have made an appointment with their primary care physicians, and can refer them to call centers if they need further assistance. This method also helps identify patients who require care management. And call centers themselves can use automated outreach, triggered by registry data, to check on patients who may need help.
Automated Program Evaluation
As management consultant Peter Drucker famously observed, "What gets measured, gets managed." Many of the tools described previously can enable an organization to measure its progress in the broad effort to manage its population's health. An advanced rules engine can integrate disparate types of data with evidence-based guidelines to show how well a healthcare system is serving various segments of its patient population.
Three additional capabilities are key to measuring progress.
First, a population health management analytics program should include a dashboard to assess the performance of individual providers and the organization as a whole. The dashboard should show how the health system or physician group measures up against metrics such as those established to receive incentive payments under pay-for-performance programs, to be recognized as a patient-centered medical home, and to achieve meaningful use of an EHR under the Health Information Technology for Economic and Clinical Health Act.
Second, the program should be able to filter population data by payer, activity center, provider, health condition, and care gaps. Managers can use these reports to spot areas of population health management that need improvement.
Third, the system should be able to generate reports that show where individual patients' health indicators remain out of bounds. If the organization's population health management processes have been unable to reach these patients, those methods may need to be fine-tuned.
A Crucial Transition
Clearly, population health management should become a core competency of all healthcare organizations as payment models change. As such, the technical capability for assessing population health needs and deploying interventions should be an integral component of a health system's technology infrastructure to reach every patient cost-effectively and provide appropriate care. Current health IT tools can perform some of this work and help providers achieve an ROI to some extent today, but providers will require additional applications to automate and support a full transition to managing population health.
Healthcare executives who look at their investment in population health management as a way to make a quick buck are barking up the wrong tree. They should recognize that the future belongs to organizations that are responsible stewards of healthcare resources. The point of investing in managing population health today is to be prepared for that new world when it arrives, instead of having to scramble to meet the challenge. If health systems take that long view and see the steps required to develop the skill sets and infrastructure-including automation capabilities-to manage their populations as an evolution, they will experience less pain and greater success in their transition to the new business model.
Michael B. Matthews is CEO, Central Virginia Health Network and MedVirginia, Richmond, Va. (firstname.lastname@example.org).
Richard Hodach, MD, PhD, MPH, is chief medical officer, Phytel, Dallas (email@example.com).
a. For convenience, we will refer to all of these people as "patients," allowing for the fact that many are not actually ill and do not require immediate care.
b. Michigan Primary Care Transformation Project, Information for Michigan POs/PHOs and Payers, Michigan Department of Community Health, Sept. 8, 2011.
c. Duncan, I., Healthcare Risk Adjustment and Predictive Modeling, ACTEX, Winsted, Conn., 2011.
d. National Quality Forum, Synthesis of Evidence Related to 20 High Priority Conditions and Environmental Scan of Performance Measures, Final Report, Jan. 13, 2010.
e. Hodach, R., ACOs and Population Health Management, How Physician Practices Must Change to Effectively Manage Patient Populations, American Medical Group Association.
Publication Date: Monday, April 02, 2012