Section 3: Selecting and Targeting Populations for a Care Management Program

An integral part of any care management program is a thorough understanding of the population it will affect. A State must decide which populations to target and how to identify and stratify members for enrollment into the program. In doing so, program staff will be better equipped to tailor appropriate interventions and resources to impact members most effectively.

Incorporating information from the 13 State Medicaid care management programs in the initial AHRQ Learning Network and additional literature, this section of the Guide, Selecting and Targeting Populations for a Care Management Program, provides information to State Medicaid staff and policymakers about:

• Selecting included populations.
• Identifying and stratifying eligible members.
• Enrolling members.

Population Selection and Enrollment

The model in Exhibit 3.1 depicts the process a State must consider in selecting members for enrollment in the care management program. States first must select the eligible population that the program will target and then determine how they will identify potential members. States have employed various techniques to identify and stratify members, including claims data analysis, physician referrals, and predictive modeling. Once eligible members have been identified, program staff must begin enrolling members. A program's enrollment strategy will depend largely on program design, including when and where to enroll members, whether to use consumer incentives, and how to retain enrollees. Careful planning during each step of the process will ensure that a program is targeting the appropriate population, identifying all eligible members, maximizing enrollment efforts, and allocating resources efficiently.

Selecting Included Populations

When designing a care management program, a State must decide which populations to include. The populations within a fee-for-service (FFS) or a primary care case management (PCCM) program vary greatly across States. States might have no full-risk managed care and, therefore, have their entire Temporary Assistance for Needy Families (TANF) and Supplemental Security Income (SSI) populations in FFS or PCCM. Other States might have only the SSI and other special populations in FFS or PCCM.

States have several options in deciding which populations to include in a program. Care management programs can target specific chronic conditions or focus on high-risk and high-cost beneficiaries. Programs can also follow a "population-based" approach by including their entire FFS and PCCM population and providing interventions appropriate for the member's risk level or disease.

Specific Diseases

In deciding on the appropriate population for their care management programs, States can choose to include members diagnosed with specific chronic diseases, the most common being asthma, diabetes, congestive heart failure (CHF), coronary artery disease (CAD), and chronic obstructive pulmonary disorder (COPD). States also target other conditions, including high-risk obstetrics and, more recently, mental health and obesity. Exhibit 3.2 shows some of the conditions targeted by care management programs in the initial 13 States involved in the AHRQ Learning Network.

Exhibit 3.2. State Chronic Disease Coverage

States	Asthma	Diabetes	CHF	CAD	COPD	High Risk Obstetrics
Arkansas						√
Illinois	√	√	√	√	√
Indiana		√	√
Iowa	√	√	√
Kansas
North Carolina	√	√	√
Oklahoma
Pennsylvania	√	√	√	√	√	√
Rhode Island
Texas	√	√	√	√	√
Virginia	√	√	√	√
Washington	√	√	√		√
Wyoming	√	√	√	√	√	√
Total	8	9	9	5	5	3

Kansas, Rhode Island, and Oklahoma do not target specific diseases through their care management program, but rather target high-risk or high-cost members. Please go to the Appendix for more information on specific States' programs. State staff also can select which conditions to manage based on additional factors, such as disease prevalence within the State and timeframe for cost savings.

Disease prevalence within the State. State staff can analyze data from State-specific or national data sources to identify the most prevalent diseases within their State. States should also consider using claims analysis to identify the most prevalent diseases specific among their Medicaid population.

North Carolina chose asthma as the first disease for care management based on a review of Medicaid claims data. Some of the key utilization factors included the following:

• In fiscal year 1998, before North Carolina implemented its program, the North Carolina Medicaid program spent more than $23 million on asthma-related care.^g
• Approximately 14 percent of the Medicaid population was diagnosed with asthma.^h
• Analysis of Medicaid claims data from the Community Care of North Carolina (CCNC) sites demonstrated that the primary and secondary reasons for both hospital admissions and emergency room (ER) visits for patients under 21 were asthma.

With these utilization statistics, North Carolina was able to show its physician workgroup that asthma constituted a serious issue within its Medicaid population, thereby building support for asthma care management.

Arkansas' care management program, Antenatal and Neonatal Guidelines, Education and Learning System (ANGELS), was established in 2002. ANGELS provides high-risk obstetrics and neonatology services through telemedicine for Arkansas women in rural areas. The State seeks to work with mothers and physicians to ensure healthy births, thereby reducing the number of medically fragile children. Arkansas Medicaid decided to create this program due to the high cost of medically fragile children, as determined through claims data. It also used statewide data to determine that Arkansas had a high incidence of low birth weight compared with the Nation.

Sources of Disease Statistics AHRQ State Snapshot Web Tool. Centers for Disease Control and Prevention (CDC)—State Data and Trends. State-Based Diabetes Prevention and Control Programs. State Heart Disease and Stroke Prevention Programs. Kaiser Family Foundation, State Health Facts.

Timeframe for cost savings. When choosing diseases, a State should consider the timeframe in which it needs outcomes. If a State requires outcomes within a short period, it should choose a set of diseases that can provide initial outcomes quickly. For example, in managing asthma, programs can expect to see outcomes and savings in a relatively short period of time compared with diabetes, which requires behavior change on the member's part and, thus, likely will fail to see substantial savings in the short term. Please go to Section 8: The Care Management Evidence Base for additional information. Below are common diseases and considerations, based on literature searches and program examples, which a State should take into account before selection.

Asthma. Asthma is a highly prevalent disease among the TANF population, but its costs are relatively low compared with other chronic conditions. Moreover, asthma is relatively easy to manage. With monitoring, proper use of medications, control of the environment, and avoidance of triggers, such as pet dandruff or second-hand smoke, most children and families can be relieved of the burden of asthma. Care management could help prevent ER visits and hospitalizations, but the savings might be lower for this disease than others because its overall costs are lower. An evaluation of North Carolina's program projected cost savings for asthma at $3.3 million in the program's first 3 years. The projected asthma savings increased every year.

Diabetes. Diabetes is a difficult disease to manage, because it requires behavior change by the member. Furthermore, because many of the outcomes of diabetes care management are seen much later, when complications (e.g., kidney failure) are avoided, diabetes management is unlikely to generate cost savings in the short term. However, evidence suggests that care management programs can reduce members' HbA1c levels and increase their compliance in getting recommended exams.ⁱ North Carolina's evaluation projected cost savings at $2.1 million in the program's first 3 years. The projected diabetes savings increased after the first year, similar to asthma. An evaluation of Indiana's program found no statistically significant cost savings for diabetes after its first 17 months. Specifically, the data showed an increase in cost among high-risk members and a decrease in cost among low-risk members. Indiana expected no cost savings at such a short time interval, based on articles in the literature on diabetes care management and its evaluator's consulting.

Congestive heart failure. Indiana's program evaluation found statistically significant cost savings for CHF in its random-control trial and time series evaluation. Members in the disease management program had lower hospital and care management services costs but higher drug costs. The net savings found were $720 PMPM, or $36 million annually, for 4,300 members statewide. An evaluation of Washington's care management program targeting CHF found no significant benefits.

Coronary artery disease. Highly prevalent among the Medicaid population, CAD is targeted consistently by care management programs. Research indicates that care management interventions for CAD can potentially reduce LDL levels and increase the use of aspirin, beta-blockers, and ACE inhibitors. Additionally, the Congressional Budget Office literature review found that most studies of care management programs for CAD reported improvements in coronary risk factors. However, improvements did not necessarily translate into lower mortality or cost-effectiveness.^j

Chronic obstructive pulmonary disease. COPD is the general term for chronic bronchitis or emphysema. Prevalent in the United States, COPD is the fourth leading cause of death.^k5 Evidence suggests that care management programs for COPD could decrease members' ER utilization and potentially save programs significant amounts of money.^l In-person care management and decision support for providers are the main interventions employed to decrease ER utilization and increase savings.

In selecting the specific diseases for care management programs, States should consider the following:

• Stakeholder Input. Stakeholders might have specific diseases that resonate with them personally. With open lines of communication, States can ensure that they are addressing these diseases, resulting in greater stakeholder support. Please go to Section 2: Engaging Stakeholders in a Care Management Program for more information on strategies to engage stakeholders.
• Available Staff and Resources. States must ensure that they have the appropriate infrastructure, including resources and qualified staff, to support interventions that impact selected diseases.
• Concurrent State Initiatives. Program staff should identify potential overlap between the new care management program and other State programs. States do not want to provide similar services to the same population, so programs might decide to cover different diseases that are not covered in an already established program.

High-Risk, High-Cost, or High-Utilization Members

While early care management programs have targeted members with specific diseases, States are moving toward programs with a more holistic approach to managing diseases. This approach addresses members' multiple physical, mental, and social needs, including comorbidities, which are highly prevalent among the Medicaid population. Members with comorbidities are typically the highest-cost beneficiaries in a Medicaid program. In fact, of the most expensive 10 percent of Medicaid beneficiaries, more than three-quarters are diagnosed with multiple chronic conditions. Additionally, nearly one-third of these members have an associated mental health disorder, such as schizophrenia, bipolar disorder, or depression.^m To manage this population more effectively, many States are beginning to target members who are at high risk for future costs or members who are currently high cost or high utilization.

Managing Comorbidities Traditionally, States have selected patients based on primary diagnosis. Although many programs are shifting to members with comorbidities, currently only a limited body of research has looked at successful interventions in managing these high-risk populations. Members with comorbidities often seek care from several specialists and require coordination among multiple providers. Disease-specific ("siloed") self-management techniques and interventions are inadequate to address the needs of patients with multiple comorbidities. Instead, programs managing complex comorbid conditions require an integrated delivery system that incorporates various strategies for addressing member needs, including behavioral health and non-medical support services.

Targeting high-risk members. States determine high-risk status by various mechanisms. Many States employ predictive modeling (which will be addressed later in this section) or claims analysis to identify high-risk members. Depending on population size, programs might target 10 percent or 20 percent of members at the highest risk. Washington, for example, uses predictive modeling to target members with risk scores in the highest 20 percent. Within this category, some programs exclude certain groups, such as patients with cancer or in long-term care facilities.

Targeting high-cost and high-utilization members. Other States have chosen to target either high-cost or high-utilization members. High-utilization members are typically among the highest-cost members and frequently visit the ER for care because of their uncontrolled condition or conditions. Programs that target high-utilization or high-cost members can identify eligible members through claims data analysis, which requires fewer resources than running predictive modeling. Many of these programs provide interventions to help patients self-manage their disease and, thus, curb inappropriate utilization of services.

As part of Oklahoma's care management program, the State created the Emergency Room Utilization Initiative to target members with inappropriate ER utilization. The initiative includes:

• PCCM provider profiling of assigned enrollees' ER utilization to show PCCM primary care providers their patients' ER utilization rates and how these rates compare to these providers' peers.
• Outreach to beneficiaries with high ER utilization (four or more visits in a quarter).
• Interventions such as letters, telephone calls, primary care provider assignment, and location of specialists.
• Followup on nurse call line calls that directed beneficiaries to the ER.

Population-Based Approach

Some States, such as North Carolina and Wyoming, have chosen to take a population-based approach to their care management programs. These States include their entire FFS and PCCM populations in the program and offer them a continuum of care based on their current needs. Members without chronic diseases might have access to a nurse call line and might be assigned to a medical home. Members who currently are able to self-manage their chronic conditions might receive educational brochures, while members who have unmanaged or poorly managed conditions (including comorbidities) are assigned to higher levels of care management. This program design permits members to move among levels of care as their conditions allow. Similar to targeting members at high risk or with high utilization patterns, the population-based approach might be more successful at managing comorbid conditions by addressing members' entire needs rather than just a single disease.

Wyoming's care management program, The Health Management Program, was implemented in 2004. All Medicaid beneficiaries are eligible to receive wellness and preventive services. Members with chronic diseases as well as depression and high-risk maternity cases may participate in the care management program. Once enrolled, members are stratified by risk level and assigned to one of three levels of intervention. Interventions range from prevention and education for members at the lowest risk level to complex case management for members at highest risk.

1. Prevention and Education. Members in this level receive a Healthwise Handbook, which is a self-care guide covering topics from ear infections to diabetes with clear, easy-to-understand information and illustrations.
2. Disease Management. Health coaches or case managers—registered nurses with specific experience in a clinical area and at least 3 years of acute care management experience—provide the majority of member support and education telephonically. For some diseases, such as mental health disorders, health coaches and care managers are licensed professionals or social workers certified to deal with mental health issues. The goals of health coaching are to empower members to better understand their illness and self-manage their condition, as well as to coordinate care between providers, the member, and the community.
3. Complex Case Management. Complex case management is reserved for members at the highest risk level. Health Management Program members are identified proactively via triage of pre-certification requests, analysis of claims and pharmacy data, or both. Types of cases appropriate for complex case management include trauma, genetic disorders, hemophilia, and cancer.

Population Exclusions

Some States also have opted to exclude certain individuals from their care management programs for several reasons. Because self-management is integral to the care management process, some States exclude Medicaid beneficiaries residing in long-term care facilities who have little control over their medication or diet. In addition, Medicaid beneficiaries in long-term care facilities should be receiving integrated care already and might not need additional care management. States also exclude waiver beneficiaries, because they already receive management through the waiver. These and other populations excluded by States participating in the initial care management Learning Network are shown in Exhibit 3.3.

Exhibit 3.3. State Inclusions and Exclusions

States	TANF FFS/PCCM	SSI FFS/PCCM	Managed Care	Dual Eligible	Long-Term Care	Waivers
Arkansas	√	√	X	X	X	X
Illinois	√	√	X	X	X	X
Indiana	√	√	X	X	X	X
Iowa	√	√	X	X	X	X
Kansas	√	√	X	X	X	X
North Carolina	√	√	X	X	X	X
Oklahoma	√	√	X	X	X	X
Pennsylvania	√	√	X	X	X	X
Rhode Island	√	√	X	√	X	X
Texas	√	√	X	X	X	X
Virginia	√	√	X	X	X	√
Washington	X	√	X	√	X	X
Wyoming	√	√	X	√	√	√

X = Excluded   √ = Included

However, some States include these populations to improve their care, achieve additional cost savings, or respond to legislative mandates.

Virginia's legislature mandated that the State include the Medicaid home- and community-based services waiver population in its care management program. Having begun to coordinate delivery of care management (or chronic condition care management) with the mental retardation/developmentally disabled (MR/DD) waiver population, Virginia has worked closely with MR Directors throughout the State to develop a protocol to avoid duplication of services to members. The State has learned early lessons from its experience with this population:

• Stakeholder involvement with the MR/DD population is critical.
• The traditional care management program initiation process does not work for individuals with MR/DD who have caregivers making their health care decisions.
• Separate reporting for the MR/DD population is important to enable the State to monitor these members independently from the rest of program members.

---

^g. North Carolina State Center for Health Statistics (SCHS). Childhood asthma in North Carolina report. Raleigh, NC. March 1999. SCHS Study No. 113.
^h. Ibid.
^i. Available at: Congressional Budget Office. An analysis of the literature on disease management programs. http://www.cbo.gov/doc.cfm?index=5909&type=0. Accessed October 22, 2007.
^j. Ibid.
^k. National Center for Health Statistics. Report of final mortality statistics Hyattsville, MD, 2003. Volume 55, No. 10.
^l. Steinel JA, Madigan EA. Resource utilization in home health chronic obstructive pulmonary disease management. Outcomes Manag 2003;7(1):23-7.
^m. Centers for Health Care Strategies. The faces of Medicaid II: Recognizing the care needs of people with multiple chronic conditions. Princeton, NJ October 2007.

---

Return to Contents
Proceed to Next Section