Addressing Racial and Ethnic Disparities in the Context of
Medicaid Managed Care: A Six-State Demonstration Project

Summary of Individual Plan Projects

Individual Health Plans:  A | B | C | D | E | F | G | H |  I  | J | K | L | M

Findings – Individual Health Plans

The experience of each of the plans will be presented in a narrative format describing the processes by which the individual projects were developed and target interventions initiated. Short-term outcomes data will be shared to the extent possible given the project’s timeframe relative to HEDIS® reporting for calendar year 2004.

Plan “A”

Baseline Data: HEDIS 2003

Initial analysis of all HEDIS measures identified several potential areas of disparity for further analysis. (Figure 1)

Figure 1. Quality of care disparities in rates for four specific HEDIS measures (D-link)

The plan identified disparities among plan members in the Comprehensive Diabetes Care measures, specifically HbA1c tests and LDL-C screening. African- Americans were significantly less likely to be tested than Caucasians. Greater than 50 percent of the plan’s population is African-American and diabetes is one of the top three diagnoses affecting plan members.

Table 5. HbA1c Testing

Table 6.  LDL-C Testing

Brief Overview of QI Project:

Prior to participation in the project, diabetes care had been selected as the next disease or condition in the plan’s Disease Management program, so the current project’s emphasis on reducing disparities was linked to ongoing initiatives in diabetes Disease Management. The plan’s QI project included collaboration among all disciplines within the plan’s Health Services, Member Services, and Provider Services Departments and community resources, to educate members and physicians on appropriate testing and diabetes management. The project involved identifying African-American plan members who had not been regularly screened or tested and implementing interventions to decrease this number. Members and their primary care physicians were notified of needed tests/screens; physicians were asked to share missing laboratory data with the plan; and case management was offered to any member identified with self-care deficits or who required additional support/assistance. The performance goals/improvement targets were to increase rates of HBA1c and LDL testing/screening by five percent.

Information Technology developed a disease management registry from specifications supplied by the QI staff. When developing the registry, the plan sought the advice of a major health plan to assist with formatting and data specifications. A review of existing “best practice” disease management programs, tools and interventions was conducted, resulting in the formation of reports and system applications that would allow staff to easily obtain member-specific information on a daily basis.

A flow chart of the plan’s disease management versus case management process was developed and used for planning and program implementation. An analysis of member, plan, and physician barriers was also completed using fishbone diagrams. In addition, plan benefits were evaluated to determine if access to services related to benefit design issues might predispose a member to inadequate care. The analysis of the underlying reasons for the disparities led to the following list:

• Possible language barriers (English is a second language for many physicians)
• Limited disease-specific education/counseling during office visits
• “No-show” to office visits or “missed opportunities” during office visits
• Insufficient knowledge of the disease process by the member
• Marginal commitment by the member
• Non-compliance to care regimen
• Unresponsive to plan outreach
• Lack of appropriate follow-up by the physician
• Inconsistent care standards
• Non-adherence to clinical practice guidelines
• Physicians: no priority/no member engagement/no incentive

Project interventions were multifaceted and included:

• Developed a diabetes registry utilizing broadened HEDIS specifications, eliminating the continuous enrollment criterion, to identify more diabetics.
• Developed Physician Profiles that listed member utilization, specialty visits, lab and pharmacy data. The Profiles were mailed to primary care physicians for review and placement into the diabetic member’s chart.
• Clinical Resource Nurses (CRNs) obtained lab results (HbA1c, LDL-C) for members with diabetes who were discharged from an acute inpatient setting. Lab results were entered into an Access database and communicated to Primary Care Providers (PCPs).
• Approved payment for Podiatry care when appropriate.
• PCPs were required to complete a diabetic flow sheet as applicable and submit it to plan for approval of services. Process allowed the Plan to obtain missing lab data and intervene with members (education, case management, care issues) when necessary.
• Established a diabetes management program (Diabetes Control Network) with Verispan. The program provides member education, monthly mailings, and instructs patients on self-management.
• Requested physicians use a standardized diabetes flow sheet for medical record documentation.
• Collaborated with a DME provider to conduct home HbA1c testing. This process also required the development of an approach to obtain and communicate results to the PCP.
• Community collaborations: A regional Diabetes Outreach Network partnered with the plan to collect data on the diabetic care and management of plan members, analyzed the data and reported the results back to the plan. The target population was plan members residing in the largest city in the service area.
• A diabetic educator from the state Organization of Diabetic Educators (ODE) collaborated with the plan to provide education and assistance to members who were in poor control.

Data sources for short-term outcomes/monitoring:

Monthly reports to monitor:

• Number of HbA1c and LDL-C tests/screens completed
• Results of tests/screens

Quarterly reports to monitor:

• Physicians non-compliant with clinical practice guidelines
• Number of members “well-controlled” compared to lab results and those in need of additional support

Data sources for long-term data analysis and outcomes/monitoring:

• HEDIS performance rate for the Comprehensive Diabetes Care measure in HbA1c test and control and LDL-C test and control
• Utilization measures: inpatient days, discharges and ER visits
• Cost measures: ER and inpatient total costs
• Member statistics: number of members in each of the severity levels
• Number of members referred to a Diabetes Self-Management Training Program (MODE)
• Number of members referred for Smoking Cessation Outreach
• Number of members referred to Weight Loss Program

Identified barriers/opportunities:

• Members included in the diabetes registry who are not diabetics
• Difficulty contacting members via telephone due to disconnections or incorrect numbers
• Level of education hinders comprehension
• Opportunity to communicate with members/providers
• Opportunity to develop useful and relevant reports/applications with IT
• Missing data on tests done at non-contracted laboratories

Findings:

The plan discovered that a large amount of diabetes laboratory data was missing from administrative files. Some of the missing data is the result of members being seen in the ER or admitted to acute care facilities where laboratory data are not submitted to the plan. In other cases, physicians are not requesting the necessary screens needed to assist in determining the member’s glycemic control status.

Plan staff worked to identify sites where members are having HbA1c or LDL-C screens performed but where results are not being sent to the plan. A list of hospitals submitting claims for these tests was compiled and the hospitals were contacted to request the data. Although the immediate target audience for this project was African- Americans, it is anticipated that all members with diabetes will benefit from outreach interventions and education.

Lessons learned:

• There are multiple approaches to developing a disease management program.
• Scheduled meetings and a commitment from each program participant is imperative.
• Maintain a project plan, assess, evaluate and revise when necessary.

Back to top

Plan “B”

Baseline Data: HEDIS 2003

Plan B focused on its African-American (non-Hispanic) and White (non-Hispanic) members ages 18 – 75 years who had been diagnosed with diabetes. The African-American (non-Hispanic) population is the plan’s largest ethnic group, with the highest rate of poor HBA1C control compared to other ethnic groups. Baseline data also reflect a poor control rate for the plan’s White (non-Hispanic) population as well. In addition to working on HBA1C control rates, the plan attempted to improve care and monitor changes across all of the Comprehensive Diabetes measures. Project performance goals/improvement targets were to improve HEDIS Comprehensive Diabetes measures by 2.5 percent overall.

Table 7.  Comprehensive Diabetes Measures

Brief Overview of QI Project:

Plan members already enrolled in the diabetes disease management program from six (6) practice sites were invited to participate in the project pilot. A decision to implement the Diabetes Navigator Program was made, based on prior plan experience with a successful Asthma Life Coach Program, best practices reported from other plans, and the results of a Diabetes Life Coach Program Pilot implemented in May, 2003. Data from the diabetes pilot program had demonstrated an improvement in HbA1C control (defined as HbA1C less than 7 percent) from “Pre Life Coach Program” (22 percent) to “Post Life Coach Program” (53 percent). Based on these findings, the Diabetes Navigator Program was implemented plan-wide for members with a diagnosis of diabetes with a special focus on the African-American diabetic population.

The Diabetes Navigator Program is a disease management program that focuses on glycemic control and includes interactive educational sessions, telephonic follow-up, social support and real-time data analysis for providers.

Program interventions were targeted at two groups: 23 physicians in the six participating group practices, and the members with diabetes. Physicians received patient status reports showing members’ status on the following parameters: HbA1c testing, dilated eye exam testing, LDL testing, and urine microalbumin testing. Additionally, physicians received professional education reminders based on the American Diabetes Association Standards of Care.

Members received a series of educational tools including a diabetes calendar and pocket diary, a program magnet with contact information, educational booklets on diabetes and heart disease, proper eating, high blood pressure, cholesterol, renal complications, practical tips for living with diabetes, and the ABC’s of diabetes and the heart. Members also received individual reminders regarding diabetes standards of care and their adherence to testing regime.

Specifically, interventions included:

• Twelve targeted mailings
• “Supermarket shopping tours” for high-risk members, where the dietician helps the member make better food selections during a typical grocery shopping trip. They are taught how to read food labels, and most importantly, how to determine which foods to avoid.
• Diabetes cooking classes, in which two dieticians prepared healthy meals and demonstrated how to modify favorite recipes to improve glycemic control. Members of the audience were encouraged to ask questions, all dishes were sampled, and subjects such as portion size and nutritional content of meals were discussed.
• Members enrolled in the diabetes DM program from six practices sites were invited to participate in the Diabetes Navigator Program.
• High risk members, as determined by A1c values, lack of routine testing, or MD referral, were contacted by project nurses on a regular basis and “coached” on development of realistic goals for improvement.
• Physicians received patient status reports showing member adherence to: HbA1c, dilated eye exams, LDL, and urine microalbumin testing.
• Physicians received educational reminders regarding American Diabetes Association (ADA) Standards of Care.
• Members received a series of educational tools and reminders regarding diabetes standards of care.

Measurement:

Table 8.  HEDIS 2004 Comprehensive Diabetes Measures

Findings:

Based on HEDIS 2004 data, the Diabetes Life Coach pilot program improved the plan’s overall HEDIS rate for the comprehensive diabetes measures by 8.7 percent. This surpassed the project’s established goal of improving the diabetes measures by 2.5 percent. Measures affected most significantly by the program included HbA1C control, and LDL-C testing and control. (Figure 2) The program produced improvements in quality of care for both African-American and White plan members; the disparity in quality of care observed at baseline was reduced for three of the measures but grew slightly larger for three other measures. Due to its success in improving care overall, the program is being expanded to reach members not included in the pilot.

Figure 2.  Changes in Three Diabetes Quality of Care Measures (D-link)

Back to top

Plan “C”

Baseline Data: HEDIS 2003

Based on baseline data analysis, Plan C elected to focus on diabetes care for its African-American population. Plan demographics reflect a high number of diabetic African-American members in the targeted area (urban), which also has a low volume of available primary care providers. The quality measures to be addressed included the HEDIS Comprehensive Diabetes measures of HbA1c and LDL-C testing.

Table 9. HbA1c Testing – HEDIS 2003

Table 10. LDL Testing – HEDIS 2003

Plan performance goals/improvement targets were to:

• increase HBA1c testing from 68.9 percent to 71 percent; and
• increase LDL-C testing from 70.6 percent to 73 percent for their African-American members with diabetes.

Brief Overview of QI Project:

The plan utilized an existing health data management system to identity trends and variations in provider care, and identify providers who were having difficulty with patients being compliant with HbA1c and LDL-C testing. The data management system compared clinical data with preprogrammed rules to determine if a patient’s treatment was consistent with standard practice guidelines. It also identified potential threats to a patient’s health and suggested treatments/interventions to ward off serious health risks. Members were risk-stratified according to the severity of their disease and the presence of comorbidities. Three severity levels were established (Level I, Level II, and Level III), and interventions were developed and implemented for each severity level (ranging from general mailings and newsletters for all members to case management for members needing intensive intervention.

In addition, the plan randomly surveyed provider offices to identify current external and internal resources being used with diabetic members, and provided educational materials to those offices. Finally, the plan partnered with a home health agency to provide home assessments and education to those members who required face- to-face intervention. A diabetes case manager contacted members in the level II and level III risk levels by phone to complete an assessment and formulate a plan of care. Members in level III also received educational mailings and were contacted by phone if an admission occurred or if they had excessive ED visits.

Interventions included:

• Using the data management system (Active Health) to send reminders to providers and members
• Identifying issues with laboratory vendors regarding laboratory data submitted to the plan
• Diabetes case management and outreach activities
• Member education, counseling, and risk factor modification (guided self-management)
• Provider education to include distribution of Diabetes Clinical Practice Guidelines and a Diabetes Mellitus Patient Checklist

Findings:

Data for 2004 reflected increased rates of HbA1c and LDL-C testing for the targeted African-American members. Rates of HbA1c testing increased from 68.95 percent to 77.61 percent and LDL-C testing increased from 70.66 percent to 80.35 percent. This increase met the plan’s overall goal of increasing HbA1c and LDL-C screening rates to 71 percent and 73 percent, respectively.

Figure 3. Changes in HbA1c Testing and LDL Testing Rates from 2003 to 2004 (D-link)

Back to top

Plan “D”

Baseline data: HEDIS 2003

After analyzing hybrid HEDIS data from 2002 and 2003, the plan identified disparities in care for its members with diabetes. African-American members demonstrated a statistically significant higher rate of poor control (HbA1c > 9.5) than Caucasian members. African-American members also had statistically significant lower rates of LDL testing and control than did White members. Disparities were also noted for Hispanic members, however, the disparities were not statistically significant.

Table 11.  HbA1c Testing - Hybrid Data

Table 12.  HbA1c Poor Control (> 9.5%) - Hybrid Data

Table 13.  LDL Testing - Hybrid Data

Table 14.  LDL Control - Hybrid Data

Understanding the potential impact of diabetes on health outcomes, the plan decided to focus efforts on addressing racial/ethnic disparities related to HbA1C and LDL control with an emphasis on appropriate nutrition. Discussions with health plan physicians at quality and other plan committees highlighted a need for physicians to have appropriate nutrition materials available for minority patients with diabetes. This was particularly true for the Hispanic population.

The project goals included:

• Increased PCP awareness of health disparities within the plan’s Medicaid population
• Improved PCP access to culturally appropriate nutrition educational information
• Enhanced member knowledge regarding diabetes risk factors within specific racial/ethnic groups
• Increased member awareness related to appropriate nutrition within cultural preferences
• Participation in community diabetes detection initiatives within the plan’s service area (specifically targeting high risk members)
• Expanding the availability of diabetes educational materials in different languages and literacy levels.

Brief Overview of QI Project:

The plan utilized a multifaceted approach to supplement the existing Diabetes Disease Management Program. The approach included educational and outreach efforts toward members with diabetes at various risk levels as well as addressing racial/ethnic issues relevant to diabetes management. Multiple plan staff participated in the development and implementation of the project. These staff included representatives from the Health and Lifestyle Management, Disease Management, Medical Services Departments, and the plan’s Quality Improvement Committees that included participating physician representatives.

The plan completed an analysis of underlying reasons for the observed disparities and concluded that they were related to cultural preferences and practices, access to appropriate resources (e.g., health care, healthy food), literacy levels, and patient/physician relationships. In addition, a literature search of “best practices" such as the HRSA/IHI Health Disparities Collaboratives, Center for Healthcare Strategies, and Management Sciences for Health (Providers Guide to Quality & Culture) was conducted.

Through these efforts, the plan developed an improved understanding of health literacy and effective messaging for at-risk groups, especially with regard to patient behavior and compliance with a medical regimen. The plan staff developed a better understanding of the cultural issues that effect physician and patient communication, patient education, and patient compliance, which allowed for the integration of culturally sensitive health education materials within their self-management and disease management programs.

Project interventions were identified by member, provider, and community activities:

Member Education

• Integrated culturally sensitive health education information into the newsletter for members with diabetes, focusing on nutrition, including appropriate food selection, examples of recipes to reflect various ethnic heritages and the importance of blood glucose monitoring.
• Developed telephone outreach program targeting high-risk Medicaid members with diabetes that included: screening reminders for HbA1c, LDL, eye exams, and nephropathy; diabetic education; resources for nutrition classes; and availability of the home glucometer delivery program with in-home training by a nurse educator.
• Sponsored a diabetes cooking class with a focus on the healthy preparation of ethnic foods.
• Created an information packet for members to include statistics related to high-risk populations, personal diabetes risk assessment, web based resources, various ethnic recipes, and diabetes management strategies.

Provider Education

• Held a diabetes management conference for provider clinical support staff that featured a presentation on nutrition, exercise and availability of the home