Population Health Management Technologies for Accountable Care

PHYTEL WHITEPAPER Shifting to Value Population Health Management Technologies for Accountable Care Authors: Richard Hodach, MD PhD MPH Karen Handmaker, MPP

Summary As population health management takes center stage in healthcare transformation, it s becoming clear that the medical neighborhood must be better organized to improve care coordination and ensure that all patients receive the right care at the right time. The hub of the medical neighborhood which includes primary care physicians, specialists, hospitals, post-acute care providers, ancillary services, social service agencies, and public health departments is the patient-centered medical home, which guides patients through the system. Health IT is the glue that connects the providers in the medical neighborhood to each other and to their patients. Care teams perform the work, not only of managing care for individual patients, but also of population health management. Besides the clinicians within a practice who provide patient care, the care team should be expanded to include all of the providers who deliver care to a particular patient. When electronically connected, this extended care team can continuously send and receive updates and can easily exchange views on a patient s condition and treatment. Whether the medical neighborhood is organized by clinically integrated networks, health information organizations, or accountable care organizations, it requires interoperable EHRs and advanced data aggregation, analytic and automation tools to manage population health effectively. The ten most effective health IT tools for population health management are as follows: Electronic health records Patient registries Health information exchange Risk stratification Automated outreach Referral tracking Patient portals Telehealth/telemedicine Remote patient monitoring Advanced population analytics These applications can be categorized as population-level solutions or patient-level applications. The core of both categories is the patient-centric registry, a frequently updated collection of patient data that drives both population-level analysis and care management. Patient registries are the central database for population health management in the medical neighborhood. Populated by clinical and administrative data, they re used for patient monitoring, patient outreach, point of care reminders, care management and other purposes. They can also be used for health risk stratification, care gap identification, quality reporting, and performance evaluation. At the patient level, population health management requires an organization to reach out to and engage all patients who have care gaps, whether or not they visit their providers. Physicians and care managers must also be alerted about those care gaps, and they must have a mechanism for intervening with patients who need routine care, as well as high-risk patients who require immediate attention. Traditional manual methods are too labor-intensive and time-consuming to do all of this consistently and comprehensively. So healthcare organizations must deploy automation tools and integrate them with registries and other data sources to make sure that patients receive appropriate services. Among the automation solutions that have been shown to be most effective in population health management are automated messaging systems for patient outreach; automated systems for alerting care teams about patient needs; and online health risk assessments, customized educational materials, and self-care recommendations. Telehealth and telemonitoring can help monitor the health status of high risk patients and can give all patients remote access to their care teams when they need it. Anyone that is involved in healthcare transformation including healthcare system and insurance executives and frontline providers should read this paper to get an in-depth view of where healthcare is heading and how it will get there. The more that healthcare professionals understand about the medical neighborhood and population health management, the faster they can move the ball forward. 2 Copyright 2013 Phytel Inc. All rights reserved. PHYTEL 11511 Luna Road, Suite 600 Dallas, TX 75234 800.559.3057 phytel.com

Contents Page 4 Introduction Page 5 What is population health management? Page 6 Population Health Management and the Medical Home Page 7 The Care Team in the PCMH and PCMN The Extended Care Team In the Medical Neighborhood Page 8 Organizing the Medical Neighborhood Page 9 Health Information Technology: The Nervous System of PHM Information-Sharing in the Medical Neighborhood Population Health Management Tools Page 10 Population Level HIT Applications Patient-Centric Registries Page 11 Claims and Financial Analytics Risk Stratification Page 13 Advanced Population Analytics Page 14 Patient Level HIT Applications Individual Engagement Page 15 Care Management Virtual and Remote Monitoring Page 16 Team-Based Care Collaboration Page 18 Conclusion Page 19 Notes PHYTEL 11511 Luna Road, Suite 600 Dallas, TX 75234 800.559.3057 phytel.com Copyright 2013 Phytel Inc. All rights reserved. 3

Introduction The U.S. health care system faces numerous challenges in adopting a population-based approach to health care delivery. This transformation will require a critical shift from the current approach, which focuses mainly on individual patients and episodes of illness, to an approach that emphasizes the health needs of an entire patient population. Providers will need to be aware of their patients interactions with other providers and health organizations (hospitals, specialists, mental health, behavioral health, and long-term care), as well as non-medical factors that affect their health and capacity to self-manage, including geography, socioeconomic status, and risky behaviors like smoking, poor nutrition, violence, and substance abuse. Population health has been defined as the health outcomes of a group of individuals, including the distribution of such outcomes within the group. 1 The term population can refer to geographic regions, such as nations or communities, but it most often describes a specific subgroup of patients. Examples include a population of patients with a specific disease (e.g., all of the diabetic patients in a practice), a group with gaps in care (e.g., all female patients without up-todate breast cancer screening), or simply, all of the people who identify Dr. Smith as their personal doctor. 2 While the term population health might be seen as implying a disassociation or a distancing from the individual patient, the opposite is true. The population health approach to care delivery strongly emphasizes patient engagement, quality improvement, and the coordination of care for individuals across care settings. Improving population health is one of the principles of the Institute for Healthcare Improvement s Triple Aim, which seeks to improve the experience of care, improve the health of populations, and reduce the per capita costs of care. 3 The Centers for Medicare and Medicaid Services (CMS), along with many private payers and healthcare organizations, regards the Triple Aim as the key goal of healthcare reform. Despite the importance of the population health approach, however, it will not be widely adopted until new financial incentives in healthcare evolve and become prevalent. A key barrier is the dominant fee-for-service payment system, which rewards healthcare providers for patient encounters and the volume and complexity of the services performed during those visits. This model discourages providers from caring for patients outside of face-to-face encounters or proactively seeking out patients with gaps in their preventive or chronic disease care. The current transition from fee-for-service to a budgeted payment model in which healthcare providers take financial and clinical responsibility for care is expected to facilitate adoption of the population health model. Additional barriers to the new approach include the lack of an infrastructure in most healthcare organizations for improving population health; the inability of most electronic health record (EHR) systems to generate the data or provide the analytics required for population health; and the fragmentation of healthcare in most communities. Because of this state of disorganization, there is little coordination of care within the medical neighborhood, which includes primary care physicians, specialists, hospitals, rehab and longterm-care facilities, home health agencies, pharmacies, labs, and imaging centers. Until these entities coordinate care and communicate, not only with each other but also with patients, social service agencies, and public health departments, the optimization of population health will remain a distant goal. For a medical neighborhood to optimize population health, it must be organized around the patient-centered medical home (PCMH), a primary care delivery model that is patient-centered, comprehensive, coordinated, accessible, and committed to quality and safety. 1 The PCMH coordinates the care of each patient across the spectrum of care settings with the help of health information technology. The objective of this paper is to explain how this can be done in the patient-centered medical neighborhood (PCMN) and associate important health IT infrastructure elements with key functions required for effective population health management. For a medical neighborhood to optimize population health, it must be organized around the patient-centered medical home (PCMH), a primary care delivery model that is patient-centered, comprehensive, coordinated, accessible, and committed to quality and safety. 1 4 Copyright 2013 Phytel Inc. All rights reserved. PHYTEL 11511 Luna Road, Suite 600 Dallas, TX 75234 800.559.3057 phytel.com

What is population health management? The definition of population health management (PHM) encompasses both the population to be managed and the approach chosen to accomplish that goal. For the purposes of this paper, we are going to use the Agency for Health Care Research and Quality s (AHRQ) definition of practice-based population health, or PBPH: We define PBPH as an approach to care that uses information on a group ( population ) of patients within a primary care practice or group of practices ( practice-based ) to improve the care and clinical outcomes of patients within that practice. 4 Patient engagement is also critically important in PHM. The Care Continuum Alliance (CCA), an association of stakeholders committed to PHM, defines population health improvement as a model featuring a physician-guided health care delivery system designed to develop and engage informed and activated patients over time to address both illness and long term health. Care Continuum Alliance members believe that managing health requires the active, integrated involvement of all health care professionals coordinated with the patient and their caregivers and families. 5 Key components of the CCA model include health risk assessments, health promotion programs, patient-centric health management goals and education, self-management interventions aimed at influencing the targeted population to make behavioral changes, and ongoing communications between patients and physicians, ancillary providers, and health plans. Care teams are also a key part of the PHM approach. Shifting the care model from an episode-based model to a person-centered population health model requires a team of providers who diligently monitor quality and outcomes, care for patients based on conditions and risk levels, and proactively manage patients who may otherwise slip through the cracks by delaying or avoiding care altogether. The ability to achieve the goals of population health management, however, only becomes possible when health information technology applications underpin and drive the fundamental activities of practices, providers and care teams. The 2010 AHRQ report first identified five domains of PHM that depend on HIT applications. Although this framework was developed to operate at a practice level, these domains all of which require the use of IT can be considered at a number of levels: patient, provider, practice, integrated health system and the medical neighborhood. The ability to achieve the goals of population health management, however, only becomes possible when health information technology applications underpin and drive the fundamental activities of practices, providers and care teams. PHYTEL 11511 Luna Road, Suite 600 Dallas, TX 75234 800.559.3057 phytel.com Copyright 2013 Phytel Inc. All rights reserved. 5

Population Health Management and the Medical Home The goals of population health management are compatible with the goals of the patientcentered medical home (PCMH). Like PHM, the PCMH is a completely different kind of care delivery model than that which most providers are used to. Both emphasize the need to proactively keep people healthy instead of just providing care when they re sick; shift the focus from acute care to preventive and chronic care; are predictive and proactive, rather than reactive; are continuous, rather than episodic; are wholeperson-oriented, rather than case-oriented; and offer care to all patients, not just those who present for care. An individual primary care practice can achieve some of these goals, but population health management as a whole requires a foundation established by the PCMH model, which becomes the hub for collaboration within the medical neighborhood. AHRQ defines the patient-centered medical home as an approach to the delivery of primary care that has the following characteristics: Patient-centered: A partnership among practitioners, patients, and their families ensures that decisions respect patients wants, needs, and preferences, and that patients have the education and support they need to make decisions and participate in their own care. Comprehensive: A team of care providers is wholly accountable for a patient s physical and mental health care needs, including prevention and wellness, acute care, and chronic care. Accessible: Patients are able to access services with shorter waiting times, after hours care, 24/7 electronic or telephone access, and strong communication through health IT innovations. Committed to quality and safety: Clinicians and staff enhance quality improvement through the use of health IT and other tools to ensure that patients and families make informed decisions about their health. Coordinated: Care is organized across all elements of the broader health care system, including specialty care, hospitals, home health care, community services and supports. 6 A fully developed PHM approach requires a number of capabilities and functions that are not yet found widely in the U.S. healthcare system, although there are many examples emerging as the number of PCMH practices, ACOs and clinically integrated networks continues to grow. These critical capabilities include an organized system of care, care teams, coordination across care settings, access to primary care, patient selfmanagement education, a focus on health behavior and lifestyle changes, and the use of linked EHRs and patient registries. 7 A fully developed PHM approach requires a number of capabilities and functions, including an organized system of care, care teams, coordination across care settings, access to primary care, patient selfmanagement education, a focus on health behavior and lifestyle changes, and the use of linked EHRs and patient registries. 6 Copyright 2013 Phytel Inc. All rights reserved. PHYTEL 11511 Luna Road, Suite 600 Dallas, TX 75234 800.559.3057 phytel.com

The Care Team in the PCMH and PCMN Perhaps the medical home s most significant contribution to population health management is the emphasis on a care team model. A team-based model of care ensures that all patients from the lowest risk level to the highest risk level are cared for in the right place, at the right time, and in the manner most appropriate for the patient. A medical home care team offers 24/7 access and alternatives to traditional face-to-face visits, including e-consults, group visits, and patient portals. The care team s responsibilities may include activities such as pre-visit planning, self-management counseling and creating complex chronic care plan for a patient with multiple illnesses To support the objectives of PHM, the care team also ensures that proactive processes are in place to manage the health care needs of all patients even those who have not scheduled visits, paying special attention to subpopulations of patients with chronic diseases, complex conditions, and behavioral health issues. These population management activities can require the sophisticated use of electronic health records and other population health management tools to identify and track cohorts of patients by risk level, adherence to care plans, appropriate medication use, and achievement of therapeutic targets. Once identified, care teams can also leverage technology-assisted tools to reach out to patients via phone, secure messaging, or email to encourage them to schedule community resources and others in the larger medical neighborhood. All of the providers caring for a patient should be regarded as part of an extended care team that is connected electronically and can continuously send and receive updates on the patient s condition. This care team should work off a single care plan that can be expanded and modified as the patient moves from one care setting to another. Further, a well-functioning medical neighborhood would feature seamless sharing of clinical information, reduced duplication and waste, enhanced continuity of care, shared decision making and strong community linkages. When all this happens, the medical neighborhood care team becomes even more patient-centered. To support the objectives of PHM, the care team also ensures that proactive processes are in place to manage the health care needs of all patients even those who have not scheduled visits, paying special attention to subpopulations of patients with chronic diseases, complex conditions, and behavioral health issues. and medications recently discharged from the hospital. High-performing care teams often include a combination of clinical and non-clinical staff. It is not uncommon for some staff to have multiple roles and for others to be embedded by a health plan or shared across multiple practice sites. Care teams, in various configurations, may include nurses, care coordinators, medical assistants, social workers, diabetes educators, nutritionists and/or health coaches who are dedicated to supporting patients as they navigate the health system and strive to achieve their care plans goals. needed appointments, refill important prescriptions or check in with their doctor following a hospital admission or emergency room visit. The Extended Care Team In the Medical Neighborhood To coordinate care and patient support beyond the walls of the medical home, HIT applications can be used by care teams to document and share information electronically and bi-directionally with providers (primary care and specialists), caregivers, hospitals, home health agencies, One compelling reason why PHM requires this close collaboration between the medical home and the medical neighborhood is that patients with complex, high-cost illnesses must be managed by multiple specialists. For example, a cardiologist may manage a patient s heart disease, but that same patient may also have advanced diabetes and emphysema requiring the attention of other specialists. The primary care physician who serves as the patient s medical home may not be able to address these conditions fully, but he or she should lead the care team that is providing holistic, comprehensive care to the patient. PHYTEL 11511 Luna Road, Suite 600 Dallas, TX 75234 800.559.3057 phytel.com Copyright 2013 Phytel Inc. All rights reserved. 7

In today s fragmented environment, however, it is highly likely that primary care and specialist providers treating these complex patients may not all be aware of each other s involvement and, even if they are aware, they may be on separate electronic medical systems and in different networks or health systems. AHRQ notes that these and the following challenges greatly complicate care coordination and the development of medical neighborhoods and effective extended care teams: No (or few) financial incentives or requirements for care coordination Limited financial integration across providers Practice norms that encourage clinicians to act in silos rather than coordinate with one another The complexity of coordination for high-need patients Patient self-referrals of which the PCMH is unaware Limited health IT infrastructure and interoperability 8 Organizing the Medical Neighborhood There are several ways to organize a medical neighborhood into a high-functioning system capable of managing populations and their health. A hospital system, a physician group, or an independent practice association can form a clinically integrated network (CIN) that adheres to a single set of clinical protocols and has the health IT infrastructure required for care coordination, care management, and patient engagement. 9 Alternatively, a health information organization (HIO) can leverage the connectivity it provides to participants by adding a layer of automation and analytic tools. 10-11 Accountable care organizations (ACOs), which are groups of providers committed to improving quality and lowering costs, can also organize the medical neighborhood into a network capable of managing population health. 12 But ACOs like CINs and HIOs--need primary care practices that follow PCMH principles to coordinate care and help patients navigate the healthcare system. In the view of many observers, the PCMH is an indispensable building block of the ACO. In an ACO, the population to be managed includes all of the people who receive their care from the providers that participate in the ACO. (For purposes of the Medicare shared savings program, beneficiaries enrolled in the ACO must receive most of their primary care from an ACO participating provider.) But many of these patients get some of their care outside the ACO, and population health management encompasses interventions that fall into non-traditional categories, such as social services, nutrition, and wellness programs. 13 The ACO s medical neighborhood includes not only the providers whom patients encounter in various care settings, but also many other actors in the community and sometimes outside of it. 14 The healthcare stakeholders that have the most experience in PHM are payers, which have long sought to influence patients and providers to improve outcomes and lower costs. Both government and private insurers and employers have engaged in various aspects of PHM, including disease management and health promotion. Recently, a number of health plans have formed partnerships with large provider organizations to form ACOs or ACO-like entities. In some cases, the plans are taking the lead in online patient engagement activities and/or are placing care coordinators in physician practices. In addition, as we explain later, many payers are providing claims data that can be extremely valuable to providers that are doing PHM. There are several ways to organize a medical neighborhood into a highfunctioning system capable of managing populations and their health. 8 Copyright 2013 Phytel Inc. All rights reserved. PHYTEL 11511 Luna Road, Suite 600 Dallas, TX 75234 800.559.3057 phytel.com

Health Information Technology: The Nervous System of PHM Comprehensive, reliable coordination of care in a medical neighborhood is impossible without health information technology. Health IT is a primary, if not the most important support tool available for medical neighborhoods, according to the AHRQ report. 15 This report further points out: Health IT has significant potential for facilitating physician efforts to coordinate patient care in the medical neighborhood Interoperable electronic health records enable PCCs [primary care clinicians] and specialists to share information, such as a patient s medical history, current problem and medication lists, diagnostic testing and laboratory results, and care plans. The AHRQ paper also notes that electronic health records need to become truly interoperable and must link with the public health data infrastructure to achieve their potential in the medical neighborhood. But this only scratches the surface of what health IT can do in population health management. With data supplied by registries and enterprise data warehouses, analytic software can be used to stratify the population by health risk, identify care gaps and set priorities for care planning. Automation tools can facilitate care management, patient outreach, health risk assessment, web-based education, remote patient monitoring, and other PHM functions. 16 In addition, ACOs and other organizations that seek to manage population health need analytics to measure performance across the entire medical neighborhood. With providers who are part of an ACO or a clinically integrated network, the challenge is to integrate data from many disparate clinical and financial systems so that analytics can be applied to it. To track care provided outside of the network, health plan claims data will also be required. 17 It should be noted that while AHRQ calls for the use of interoperable EHRs, systems from different vendors are unlikely to be capable of exchanging discrete data directly in the foreseeable future. Instead, healthcare organizations and HIEs create data liquidity by aggregating and normalizing data from disparate EHRs which is another form of interoperability. Information-Sharing in the Medical Neighborhood A wide range of health IT capabilities are needed to knit together the medical neighborhood in ways that facilitate PHM. This starts with the ability of providers to communicate with one another electronically. One way to do that is through health information exchanges (HIEs) that connect providers across communities. Another is to use the direct secure messaging protocol to transfer clinical data back and forth between providers. Both approaches have drawbacks that are discussed later. Beyond the sharing of information among providers and between providers and patients, PHM requires analytic and automation tools for care managers to assess the health status of patients, collaborate on care planning, and engage patients more fully in their own care. Moreover, when an ACO or a clinically integrated network takes financial risk for care, it must have tools for evaluating both clinical and financial performance. 18 Population Health Management Tools The commercially available PHM solutions described below can be categorized as population or patient level functions to emphasize new and enhanced activities required to fully implement population health management. Population level tools integrate multiple data sources, apply evidence-based and predictive modeling algorithms, and generate actionable performance reporting. Patient level tools use the output from population level applications to inform patient engagement and care team workflows. In most cases these applications are complementary to electronic health records (EHRs), which, by and large, are not designed for PHM. 19 EHRs, practice management systems, and other clinical and administrative systems can provide much of the structured data required for PHM. But as we ll see, other data sources and capabilities must also be included. Comprehensive, reliable coordination of care in a medical neighborhood is impossible without health information technology. PHYTEL 11511 Luna Road, Suite 600 Dallas, TX 75234 800.559.3057 phytel.com Copyright 2013 Phytel Inc. All rights reserved. 9

Population Level HIT Applications Patient-Centric Registries Traditionally, a registry is a list of all the patients in a physician s practice who share a characteristic, such as a certain condition (diabetes, asthma, hypertension) or medication regimen. As HIT capabilities have improved, registries are now built to be patient-centric providing a full view of all information associated with a single patient but the information can be filtered by one or many criteria such as diagnosis, medications, age, payer, lab results and more. Dynamic registries help the care team keep better track of these patients by providing care reminders and by identifying patients who are overdue for certain kinds of care, or who are not adhering to care plans. Registries can be integrated with other tools such as automated messaging systems that remind appropriate patients to schedule appointments with their provider. Patient registries form the central database for PHM in the medical neighborhood. Populated by clinical and administrative data, they are used for patient monitoring, patient outreach, point-of-care reminders, care management, public health reporting, and other purposes. When combined with analytic tools, they can be used for health risk stratification, care gap identification, quality reporting, and performance evaluation. Registries can also provide feedback to physicians to benchmark their own performance and support their continuous improvement efforts. And a registry can be the online platform that allows all providers caring for the same patient to collaborate and coordinate care across the medical neighborhood. Figure Managing A. Technologies Populations, for Population Maximizing Health Technology Management in a PCMH-N Ten Recommended Health IT Tools to Achieve PHM: Electronic health records Patient Registries Population in the Community Patient Engagement Automated Outreach Patient Portals Care Management Clinical Analytics Clinical Decision Support Patient-Centered Registry Patient Population of the Primary Care Office Health Information Exchange Payer Risk Stratification Automated Outreach Referral Tracking Claims and Cost Risk Stratification Primary Care Office Care of a patient Patient Portals Telehealth/Telemedicine Others who supply/require information and coordination Remote Patient Monitoring Advanced Population Analytics Specialty Care Hospital Radiology, Lab, Rx Distance Monitoring Referral Tracking/HIEs Telehealth/ Telemedicine Remote Patient Monitoring 10 Copyright 2013 Phytel Inc. All rights reserved. PHYTEL 11511 Luna Road, Suite 600 Dallas, TX 75234 800.559.3057 phytel.com

5% of the population accounts for nearly half of all health costs, but only 30% of the patients in the high-risk category were high cost a year prior. A registry lists each patient s demographic characteristics, diagnoses, lab values, medications, and other pertinent data. Applied to the individual patient, a registry can show when the person was last seen, who provided what care to that patient, the patient s current health status, and when the patient is due to visit again. Applied to a population, the registry can show, for example, how all of a particular provider s patients with type 2 diabetes are doing, which diabetic patients are out of control, or how well an entire organization is treating patients with that condition. Analytic applications can compare the data in a registry with nationally recognized clinical protocols or guidelines accepted by an organization s providers. Such tools can identify care gaps, help stratify the population by health risk, and generate outreach messages to patients in need of preventive or chronic care. These analytics can also be used to generate reports on subpopulations, such as patients with uncontrolled hypertension, and to alert providers and care managers that particular patients need attention. In addition, care teams can leverage registry reports to prioritize interventions with high-risk patients, create pre-visit care plans, and customize educational materials to patients in certain categories. A registry that draws its data exclusively from a single practice or healthcare organization is inherently limited because it includes only information generated by that entity. A more effective registry would contain data from all of the providers caring for each patient. However, the owner of the registry would have to aggregate the information from these providers and then normalize all of the data to a single format so that it can be displayed and analyzed. It s also essential to scrub and validate this data to assure its integrity before it is used to manage care or evaluate performance. Claims and Financial Analytics One way for a healthcare organization to obtain information on patients who receive care from outside provider networks is to obtain paid claims data from health plans. Health plans are increasingly finding ways to share this data with providers to help them reduce variations in care and manage population health. Some payers, as mentioned earlier, are even collaborating with providers to form ACOs. 20 Claims data is less actionable than clinical data because it can lag the date when services were provided to patients by a month or more. It also has deficiencies because the information is based on payment and excludes clinical lab results and other pertinent information captured in the clinical record. For example, a claim for a test ordered to rule out a diagnosis might include that diagnosis on a claim, but the patient might not have that condition. Claims information can also help an organization calculate the total cost of providing particular kinds of care. A practice management system or a hospital financial system is not designed to furnish that kind of information. But claims data, when combined with analytics, can supply an approximation of care delivery costs. ACOs and other organizations that are at financial risk must have these figures in order to stay within a budget and determine which providers are most cost effective. 21 If an organization does not have access to paid claims data, it may use billing data from its financial system as a substitute. In an ACO or a clinically integrated network, this pre-adjudicated claims data may be drawn from the billing systems of all participants. When coupled with lab values, prescription fill information and other clinical elements, the combined data set can be used to populate registries, even without EHR data or with data only from the dominant EHRs in the network. Risk Stratification With 5% of the population accounting for nearly half of all health costs, 22 it s critically important for healthcare organizations that are taking financial risk to know who those patients are. In addition, only 30% of the patients in the high-risk category were that sick a year earlier. 23 So organizations need a method to stratify their population by health risk and provide the appropriate interventions to prevent people who are moderately sick from becoming severely ill. They must also identify specific care gaps in their population to ensure that patients receive the preventive and chronic care they need to maintain their health and control their conditions. Risk stratification refers to the periodic and systematic assessment of each patient s health risk status, using criteria from multiple sources to develop a personalized care plan. A patient s health status may be reflected by a score or placement in a specific category, based on the most current information PHYTEL 11511 Luna Road, Suite 600 Dallas, TX 75234 800.559.3057 phytel.com Copyright 2013 Phytel Inc. All rights reserved. 11

available. 24 The identification of a patient s health risk category is the first step towards planning, developing and implementing a personalized patient care plan by the care team, in collaboration with the patient. For some, the plan may address a need for more robust care coordination with other providers, intensive care management, or collaboration with community resources. The care team s observations also play a vital role. The more variables included in determining the risk category, the more reliable and accurate the prediction of future health risks and costs can be. Severity of medical condition has historically been the primary factor for stratification. It is the most readily identifiable and perhaps the most useful in and of itself. However, meshing severity data with patient specific characteristics related to co-occurring medical and behavioral health disorders, patient confidence, and psychosocial risk factors such as living alone or low income, allows for a much more refined approach to stratification and informed patient-centric care management strategies and interventions. Stratification profiles can also identify patients at high risk for poor compliance or untoward outcomes that do not necessarily meet high severity criteria; conversely, patients with an illness or symptoms classified as high severity may be assigned to a relatively low risk level if the patient has other characteristics which suggest that this single risk factor is not sufficient to substantiate assignment to a high risk category. For example, a patient with a history of elevated HbA1c who has well controlled blood pressure and lipid levels and takes medications as prescribed would not be regarded as high risk. Multifactorial stratification can also be used to determine the resources required to address risk reduction across the organization and the most appropriate allocation of those resources. For instance, nurse care management might make more sense than an increase in PCP visit frequency, depending on the patient s particular profile. Thus, risk stratification can guide the timeliness of responses required by specific subgroups rather than applying one standard to all patients. Organizations can use multiple means to risk-stratify a population. Patients or employees can be asked to fill out a health risk assessment that may be online. If an organization has a registry, it can apply analytics to that database to identify patients whose health indicators suggest that they are high risk. The best approach is to do both and to pull in any other available data that will help the organization get an accurate picture of a person s total health situation. Some patients who may not yet be high risk could easily move into that category because of psychosocial factors such as living alone or having a low income or poor access to care. In addition, the risk stratification can help care managers determine the priority they assign to their patients and the types of interventions that are appropriate for each one. For example, high-risk patients may need more personalized attention, whereas those in the medium-risk category may only require automated messaging and online To evaluate its performance, an organization that engages in PHM can leverage a data warehouse or population analytic tools needed to convert its data into useful intelligence. 12 Copyright 2013 Phytel Inc. All rights reserved. PHYTEL 11511 Luna Road, Suite 600 Dallas, TX 75234 800.559.3057 phytel.com

To truly manage financial risk, organizations must learn how to manage population health effectively. And to do that, they must fundamentally change the process of care delivery, using the appropriate automation and analytic tools. education. Low-risk patients may simply be prompted to maintain their health and get appropriate preventive care. To identify care gaps, clinical analytic tools automatically and continuously apply decision support rules to the structured data in a registry or data warehouse. These analytics can generate exception reports, which identify patients who do not meet specific criteria for best practices. Examples include patients with persistent asthma who have not been prescribed an inhaled corticosteroid, patients over 50 who have not had a colonoscopy in the past 10 years and elderly individuals with multiple chronic conditions and recent hospitalizations or emergency room visits. In addition, the system should be able to generate panel reports, which show providers and care teams key indicators about all of the patients they are responsible for. These brief summaries facilitate treatment review and care planning ahead of patient encounters. Advanced Population Analytics To manage population health, a provider organization or an ACO must measure its clinical and financial outcomes. The organization must track the health status of its patient population and particularly its high-risk patients to reduce the per capita costs of care and improve the health of populations. It can also use historical data to predict what costs will be going forward. To evaluate its performance, an organization that engages in PHM can leverage a data warehouse or population analytic tools needed to convert its data into useful intelligence. Besides the patient-level reports described above, these reports must be able to give providers, care managers, and organizational management views of how well the population is being managed at a variety of levels. At the top level, for example, managers should be able to see the prevalence of common chronic diseases in the organization s population. They should be able to risk-stratify that population by condition and see how the portion of highrisk diabetic or hypertensive patients is changing over time. Managers should also be able to look at which segments of the population are generating the highest costs and how that changes, so they can shift resources as needed. And their analytic tools must enable them to evaluate the performance of individual providers and practices on both quality and efficiency. 43 A physician should be able to access reports on his or her own population of patients with a particular chronic condition and see how those patients are doing over time. These reports can help identify patients who are outliers in terms of health status and those who have not received appropriate care. Providers should also be able to compare their own performance on quality measures with national benchmarks and with the average for their practice or organization. To assess population health at a particular point in time, organizations can use measures that describe care processes, such as how many patients with diabetes received an annual eye exam, intermediate outcomes such as blood pressure or HbA1c, and long-term outcomes. The latter measures include both clinical data and patient-reported data, such as functional status and self-perceived health. Organizations must also continuously measure patient satisfaction. By tracking progress on all of these metrics over time, they can see whether they are improving the health of their population. Furthermore, organizations that are taking financial risk need to have the ability to understand how the quality of care impacts the cost of care and how that is likely to affect future costs. To do this, they must risk stratify the population as described above, but must also focus on which individual patients and population segments are likely to generate future costs. Then interventions should be designed to provide patientcentered care and shared decision making with those patients to curb risk factors and control chronic conditions. In addition, organizations must determine which providers utilize the most resources for particular types of care, risk-adjusting that data for the relative illness and compliance of their patients. With this information, they can construct reliable cost and quality profiles and steer patients to the most efficient, high-quality providers. By using risk stratification, predictive modeling and provider utilization data, organizations can obtain a perspective on their current and future costs. But to truly manage that financial risk, they must learn how to manage population health effectively. And to do that, they must fundamentally change the process of care delivery, using the appropriate automation and analytic tools. PHYTEL 11511 Luna Road, Suite 600 Dallas, TX 75234 800.559.3057 phytel.com Copyright 2013 Phytel Inc. All rights reserved. 13

Patient Level HIT Applications Individual Engagement A population health management approach requires an organization to reach out to and engage patients who have care gaps to alert them that they need to make appointments with their providers or take other action to close these gaps. Automation is crucial to systematizing this approach. Manual methods are too expensive and time consuming to ensure periodic outreach to an entire patient population, including people who have not sought care but need it. An electronic registry populated with EHR and administrative data can be the basis of this kind of outreach. When combined with evidence-based clinical protocols and analytic software, the registry can supply the information that tells the automated messaging system when to call patients who are due for particular services. Automated communications can also be used in a variety of patient education and engagement activities. A study of automated phone messaging to patients with diabetes and other patients with hypertension showed that it was effective in encouraging many of them to seek appropriate care for their conditions, including office visits and tests. 33 Similarly, the use of automated messaging to promote adherence to statin medications was shown to be effective in a large scale, randomized trial at Kaiser Permanente. 34 The same kind of automation tool that triggers appropriate and timely messaging to patients can also be used to alert physicians and care managers that particular patients need particular services or urgent interventions. An EHR can provide some of these alerts, but only when it s accessed, which would normally be during a patient visit. These kinds of reminders limited to a narrow range of health maintenance, decision support, and chronic care alerts are insufficient for population health management. 14 Copyright 2013 Phytel Inc. All rights reserved. PHYTEL 11511 Luna Road, Suite 600 Dallas, TX 75234 800.559.3057 phytel.com

Once an organization has reorganized its care processes within the context of the medical neighborhood, a subsequent step in its journey to population health management is to use technologies designed to provide care and patient education remotely. Care Management One of the hallmarks of an organization prepared to manage population health is the use of care managers to coordinate care and ensure that all patients receive appropriate interventions. As they do in traditional disease management programs, these care managers work closely with high-risk patients to reduce exacerbations of their conditions that can lead to ER visits or hospitalizations. But to optimize population health, they must also maintain contact with other patients who fall into the low-risk and medium-risk categories. It is impossible for them to manage the care of so many patients without the use of automation tools. A few years ago, researchers calculated that it would take 18 hours a day for a primary care physician to provide all evidencebased preventive and chronic care to a typical panel of 2,300 patients. 25 Just to deliver all recommended care to a panel of 2,500 patients for the 10 most common chronic diseases would take more than 10 hours per day, another study found. 26 While there is no comparable published research on the amount of time it takes care teams to do that work, unpublished data from a large Midwestern group indicates that care management requires an average of 138 minutes of staff time per patient. When that figure is compared to the prevalence of complex chronic conditions in a typical primary care primary care practice, it can be inferred that a single PCP with a panel of 2,500 patients would require 1.35 FTE care managers, and a 10-doctor practice would need 13 care managers. 27 Most physician practices cannot afford so many care managers. Moreover, many care management tasks are routine and do not require the involvement of clinicians. Automation can perform these routine jobs, freeing nurses and doctors to care promptly for the patients who need their attention. Automated patient outreach and the automatic alerting of care teams to patient care gaps are examples of such an approach. In addition, health risk assessments can be offered online; assessments, customized learning materials, and self-care recommendations can be sent to patients via web portals or secure e-mail; and campaigns can be designed to improve the care of all relevant subpopulations. When applied to patients with chronic diseases, these campaigns can be tailored for people in hundreds of different subcategories. For example, a diabetes population can be classified into patients with type 1 diabetes, type 2 diabetes, type 2 diabetes and hypertension, poorly controlled type 2 diabetes, and so forth. Different educational materials and self-care recommendations would be sent to patient populations in each cohort, using one or more automated modes of communication, including text, phone and email, as preferred by the patient. Virtual and Remote Monitoring Once an organization has reorganized its care processes within the context of the medical neighborhood, a subsequent step in its journey to population health management is to use technologies designed to provide care and patient education remotely. These applications fall roughly into three categories: telehealth, which automates the process of keeping track of changes in patient health status; telemedicine, which permits patients to consult with physicians or nurses through audio and video conferencing; and web portals, which can be used to share information and interact with patients online. Telehealth can include home monitoring, mobile monitoring or a combination of the two. Home monitoring is used most frequently with high-risk patients such as those with congestive heart failure or people recovering from operations; but it has also been used successfully to help people control other chronic conditions such as diabetes and hypertension. Monitoring data can help inform care plans and can form the basis for automated or live feedback to patients on their health management. A number of mobile apps designed for smartphones and tablets also enable consumers to monitor their own conditions. The value of most of these apps has been limited so far, because few physicians are viewing the monitoring data. 28 That s expected to change, however, as provider organizations increasingly take financial risk for care delivery. Telehealth provides continuous data on a patient s condition and increases patient engagement in their own care. Studies show that this can improve patient outcomes and that it is cost effective. 29 But for healthcare organizations to make good use of this data in care management, they must activate automated protocols so that care managers and physicians are alerted only PHYTEL 11511 Luna Road, Suite 600 Dallas, TX 75234 800.559.3057 phytel.com Copyright 2013 Phytel Inc. All rights reserved. 15