Executive Summary

Project Overview

The goal of the Comparative Effectiveness of Health Care Delivery Systems for American Indians and Alaska Natives Using Enhanced Data Infrastructure project, also referred to as the IHS Improving Health Care Delivery Data Project,^a was to create an Indian Health Service (IHS) data infrastructure to provide information on health status, utilization, and treatment costs to inform the identification and prioritization of effective strategies for chronic disease management and to improve health outcomes among American Indians and Alaska Native peoples (AI/ANs). The project data were used specifically to assess strategies for providing health services for AI/ANs with diabetes or cardiovascular disease (CVD) through two objectives:

Develop a longitudinal data infrastructure from existing electronic data stored on multiple platforms to provide information about AI/AN health status and IHS health service utilization and treatment costs.
Conduct a comparative effectiveness research (CER) study on strategies implemented to reduce CVD risk among AI/ANs with diabetes or CVD using the data infrastructure.

The project was funded by the ��Ƶ (AHRQ) with in-kind support provided by IHS and Tribal health organizations. The Centers for American Indian and Alaska Native Health (CAIANH) at the Colorado School of Public Health, University of Colorado Denver implemented the project.^b The primary source of data for the infrastructure was the IHS National Data Warehouse.

IHS funds health services for approximately two million AI/ANs. Similar to other health systems, IHS and Tribes have implemented many initiatives to address chronic disease. For example, IHS is incorporating patient-centered, medical home concepts throughout its system, using a model referred to as the Improving Patient Care (IPC) Program.¹ IPC sites are improving the quality of, access to, and coordination of services by focusing on patient- and family-centered care; ensuring access to primary care by health care teams; acting on the guidance of the community and of Tribal leadership; and making positive, sustainable, and measurable improvements in care.¹ Additionally, the congressionally funded Special Diabetes Program for Indians (SDPI) provides resources to IHS and Tribes to prevent and treat diabetes, and to monitor health outcomes for those with diabetes.² SDPI provides grants to over 400 IHS, Tribal, and Urban Indian health programs.² Using current scientific research and evidence-based best practices, SDPI grant programs have made tremendous improvements in diabetes treatment and prevention in both clinical settings and community-based programs. Since SDPI was implemented in 1998, access to education and case management (ECM) services by those with diabetes has increased dramatically.² At the same time, there have been documented improvements in glycemic control (HbA1c) and low-density lipoprotein (LDL) cholesterol levels among AI/ANs with diabetes, as well as decreases in the incidence of end stage renal disease (ESRD).²

The IHS Improving Health Care Delivery Data Project was designed to provide information on the prevalence of comorbidities, health service utilization, and treatment costs for AI/ANs with diabetes or CVD. The fiscal year (FY) 2010 project population included persons who lived in 14 project sites, nearly 30 percent of all FY2010 IHS active users.^c This was the first time that such data were available for a large percentage of the population served by IHS. Key FY2010 findings include:

14.7 percent of adults (i.e., persons aged 18 years and older) had diabetes and, among these adults, 31.6 percent also had CVD;
5.4 percent of adults had CVD but not diabetes;
Among persons with diabetes, the average number of hospital emergency department visits was one, as was the average number of days spent in the hospital;
30.8 percent of inpatient stays by adults with diabetes or CVD, that were not for obstetrical care, were found to be ambulatory sensitive; that is, they may have been prevented with access to and use of outpatient services;
Among adults with diabetes, the average number of primary/general clinic visits, including general office and diabetes clinic visits, was 5.8; the number for those with CVD but not diabetes was similar;
41.1 percent of adults with diabetes had at least one ECM visit; among these adults, the average number of visits was 2.8;
Adults with CVD but not diabetes, as compared to those with diabetes, were less likely to use ECM services; however, those that did had a greater number of visits; and
Average IHS total treatment costs for adults with diabetes were estimated to be $8,164. Treatment costs for adults with diabetes but not CVD (i.e., $6,129) were approximately half the costs for those with both conditions ($12,568). The 14.7 percent of adults with diabetes accounted for one-third of adult treatment costs.
Among adults with CVD but not diabetes, estimated treatment costs averaged $9,009.

IHS has implemented a number of initiatives to meet the needs of AI/ANs with chronic disease. It is hoped that IHS and the project sites may use these findings to identify additional opportunities to enhance services for those with diabetes or CVD to improve their health outcomes and ensure effective use of IHS and Tribal (I/T) health resources. For example, data on use of primary care services, including ECM and home services, may inform programs and policies aimed at improving access to and use of such services, as well as coordination among service providers. Below is an overview of project methods and key findings, followed by a discussion that includes a description of project limitations.

Methods

The data infrastructure includes data for a purposeful sample of approximately 540,000 AI/ANs who obtained health services through IHS and lived in 14 Service Units. At least one of the 14 project sites (i.e., Service Units) is located in each of the 12 IHS Areas: Aberdeen, Alaska, Albuquerque, Bemidji, Billings, California, Nashville, Navajo, Oklahoma City, Phoenix, Portland, and Tucson. Because of the large number of AI/ANs living in the Navajo and Oklahoma City Areas, two project sites were selected from these areas, for a total of 14 project sites. In nine of the project sites, IHS provides services; the remaining five project sites were Tribally operated.

The data mart includes four consecutive years of data (FY2007-2010) from three primary sources:

IHS National Data Warehouse (NDW): Information on utilization of I/T inpatient, outpatient, and pharmacy services.
Contract Health Services (CHS): Utilization and payment information for services obtained at non-IHS providers^d yet paid for by IHS.
Centers for Medicare and Medicaid Services Cost Reports: Information on the costs of providing I/T services within the IHS Service Units.

Each type of data was stored in different computer platforms. CAIANH obtained data extracts from each source and created a series of analytic files for the data infrastructure. Statistical software was used to merge and analyze the data, and other nationally recognized software was used to enrich the data by creating additional project measures. A nationally recognized risk-adjustment software program (Risksmart™) was used to identify conditions for which patients were treated, by categorizing the diagnostic codes on the utilization records into meaningful condition categories, and to assign patients a health risk score based on their age, gender, and acute and chronic conditions. Patients with more high-cost acute and chronic conditions had higher risk than those without such conditions. In addition, ��Ƶsoftware was used to identify hospital inpatient admissions that were sensitive to ambulatory services; that is, admissions that may have been prevented with access to and use of outpatient services.

Although IPC was not fully implemented in FY2010, one IPC aim is to provide coordinated primary care services. Primary care services, as compared to specialty care services, include services provided during general office or diabetes clinic visits with physicians, physician assistants, and nurse practitioners; ECM visits by nutritionists, nurse educators, case managers, and clinical pharmacy specialists; and home visits. For this project, we define ECM services as visits conducted specifically for ECM, while recognizing that education and case management are also provided during general primary care and diabetes clinic visits, and by telephone.

One of the goals of this project was to evaluate the provision of ECM services. Due to issues related to data availability and the identification of ECM services, CAIANH was not able to fully implement the CER study during the AHRQ-funded project period. However, two important tasks for the CER study were completed. First, an algorithm was developed to identify five different types of ECM services. Second, analyses were conducted to examine patient characteristics associated with use of ECM services among adults with diabetes or CVD.

The first task involved the identification of types of I/T-provided ECM visits, since provision of ECM services varied across the project sites, as did documentation of provided services. In collaboration with the project’s Steering and Health Information^e Committees, an algorithm was developed to identify five types of ECM visits: 1) diabetes education;^f 2) nutrition education; 3) clinical pharmacy (i.e., visits conducted by clinical pharmacy specialists who provide services using an advanced practice pharmacy model;³ 4) case management, and 5) other types of health education (e.g., obesity, smoking cessation). It is important to note that ECM visits in categories 2-4 may have been provided for individuals or patient groups in a diabetes education clinic. For this reason, we also report utilization of diabetes education clinic services regardless of provider type.

NDW and Cost Report data were used to estimate costs associated with providing I/T services. IHS total treatment costs include those associated with I/T and CHS services.

The project’s success was facilitated by close collaboration among the IHS Division of Diabetes Treatment and Prevention, Pharmacy Program, and Office of Information Technology; the project’s Steering and Health Information Committees; and CAIANH. The committees provided advice and guidance on project implementation, the creation of data measures and project reports, and interpretation of project findings. The Health Information Committee was comprised of representatives from each project site. Collaboration with the project sites was further facilitated by an average of three site visits to each location. In addition to committee Webinars, email and phone communication, and the site visits, CAIANH conducted a training program to provide technical assistance on the development of the data infrastructure and options for analyzing project data, including training on SPSS. Representatives from the IHS Pharmacy Program, IHS Division of Diabetes Treatment and Prevention, and five project sites attended the training. At the conclusion of the project, copies of the data infrastructure were provided to IHS. For project sites that requested them, copies of data infrastructure analytic files for their project site’s population were provided.

Key Findings

Although the data infrastructure includes information for four fiscal years, this report includes FY2010 findings. The FY2010 project population included 437,608 IHS active users who had their community of residence in one of the 14 project sites. An active user is a person who used services at least once during the past three fiscal years. For example, an FY2010 active user obtained services at least once during FY2008-FY2010.

Health Status

Three different types of measures were used to describe the health status of the project population. They included 1) the prevalence of specific conditions; 2) health risk; and 3) clinical measures of glycemic, blood pressure, and cholesterol control.

In FY2010, 14.7 percent of AI/AN adults had diabetes. Among these adults, 31.6 percent also had CVD. The prevalence of hypertension, renal disease or failure, treatment associated with amputations, mental health disorders, and liver disease among the adults with diabetes was 77.9 percent, 13.3 percent, 2.4 percent, 23.5 percent, and 6.8 percent, respectively. A number of adults had CVD but did not have any record indicating they had diabetes. Of all adults, 5.4 percent had CVD but not diabetes.

Each person was assigned a health risk score to reflect his or her morbidity burden and expected use of health resources. The health risk score is benchmarked to a U.S. commercially insured population, which had an average health risk of 1.0. The average health risk of the FY2010 AI/AN project population was 1.3, or 30 percent higher than the average risk for the U.S. commercially insured population. The average health risk of AI/ANs with diabetes was more than four times higher than that of the reference population (i.e., 4.6), while that of AI/ANs with both diabetes and CVD was nearly nine times higher (i.e., 8.8).

Although optimal levels for glycemic control (HbA1c), blood pressure, and LDL cholesterol are determined by a patient’s health status, age, and other factors, IHS general treatment guidelines were used to report on these clinical measures. Approximately 63 percent of adults with diabetes who had an HbA1c test result had HbA1c values less than 8.0 percent. Nearly 71 percent of adults with diabetes and a blood pressure result were found to have systolic blood pressure values less than 140 mmHg, and 65 percent of those with diabetes and a cholesterol result were found to have LDL cholesterol values less than 100 mg/dl. Among adults with diabetes who had data for all three clinical measures, 31.1 percent had values within these ranges for all three measures.

Health Service Utilization

Hospital inpatient service utilization includes use of I/T services and CHS services. In FY2010, the hospital admission rate for AI/ANs with diabetes was 0.23 admissions per person, and AI/ANs with diabetes spent, on average, one day in the hospital. AI/ANs with CVD but not diabetes averaged 1.3 hospital inpatient days during the year. AI/AN adults with diabetes or CVD accounted for over 60 percent of all I/T adult hospital stays, excluding those for obstetrical care. Among these inpatient stays, 30.8 percent were classified as sensitive to ambulatory services. Utilization of I/T hospital emergency department (ED) services by AI/ANs with diabetes averaged 0.9 visits during FY2010; ED utilization of those with CVD but not diabetes was similar.

Among AI/ANs with diabetes, the average number of general primary and diabetes clinic visits^g was 5.8 during FY2010. The number for those with CVD but not diabetes was similar. The average number of ECM visits among all persons with diabetes was 1.2; their average number of home visits was 0.4. More information on ECM services used by adults is provided below.

IHS Total Treatment Costs

During FY2010, the average cost of treating an adult with diabetes was estimated to be $8,164. IHS total treatment costs for adults with diabetes but not CVD ($6,129) were approximately half the costs of those with both conditions ($12,568). Although the prevalence of diabetes among the AI/AN adults was just under 15 percent, their treatment costs represented one-third of all treatment costs for adults. Among adults with CVD but not diabetes, estimated treatment costs averaged $9,009.

Factors Influencing ECM Utilization

Due to issues related to data availability and the identification of ECM services, CAIANH was not able to fully implement the CER study during the AHRQ-funded project period. However, two important tasks for the CER study were completed. First, an algorithm was developed to identify five types of ECM services. Second, analyses were conducted to examine patient characteristics associated with use of ECM services among adults with diabetes or CVD. Findings from this work are presented below.

Just over 40 percent of AI/AN adults with diabetes utilized ECM services during FY2010. The average number of ECM visits by adults with diabetes who had at least one visit was 2.8; 42.3 percent had one visit and 57.7 percent had two or more. Of all ECM visits by adults with diabetes, 36.7 percent were classified as clinical pharmacy visits, 33.1 percent were for nutrition education, 17.0 percent were for diabetes education, 3.8 percent were for case management, and 9.3 percent were for other types of health education (e.g., obesity, smoking cessation). Diabetes education clinic visits, regardless of provider type, accounted for 38.9 percent of all ECM visits by adults with diabetes.

ECM utilization patterns for AI/ANs with CVD but not diabetes differed somewhat from those of AI/ANs with diabetes. Among adults with CVD but not diabetes, only 10.6 percent used ECM services. However, among those that did, the average number of ECM visits was 4.5. The majority of these visits were provided by clinical pharmacy specialists.

In addition to health status, findings from a logistic regression model indicated that third-party health coverage (e.g., Medicaid, Medicare, private) and use of other primary care services (i.e., primary/general office and diabetes clinic visits) influenced ECM utilization among adults with diabetes or CVD. This information may be used to assess options for improving access to and use of ECM services. For example, ECM utilization results may be used to assess the provision of ECM services for adults with multiple chronic conditions, with Medicaid or no health coverage, and with low use of other primary care services. In addition, the findings will inform future work on the CER study, for which CAIANH will examine the influence of ECM utilization, by provider type, on patient treatment costs using statistical models that control for issues related to selection bias in observational studies.

Discussion

The IHS Improving Health Care Delivery Data Project provided IHS and the collaborating project sites critical information on the health status, service utilization, and treatment costs for close to 30 percent of the FY2010 IHS active user population.⁴ Previous to this project, IHS had such data for only one Service Unit.^5,6

Project findings may inform IHS and Tribes in the identification and prioritization of effective strategies for chronic disease management. For example, information on comorbidities among those with diabetes may facilitate the provision of services to better meet their needs. Findings concerning hospital ED and inpatient service utilization may be used to identify opportunities to improve access to and use of outpatient services. Data on the number of office visits, ECM visits, and home services obtained by persons with diabetes or CVD may be used to guide efforts to enhance the provision and coordination of these services, which may ultimately reduce unnecessary use of hospital services. Results concerning the percentage of patients with diabetes or CVD who use ECM services, the number of visits they had, and visit type may be used to assess options for providing and improving access to the appropriate level of service. Finally, the treatment cost findings will increase understanding of existing patterns of health resource allocation.

Health coverage results indicate that approximately half of AI/ANs who used IHS services had no third-party coverage in FY2010. Among adults with diabetes or CVD, nearly 40 percent had no third-party coverage. Implementation of the Patient Protection and Affordable Care Act will provide AI/ANs increased opportunities to obtain health coverage, and I/T providers additional opportunities to obtain reimbursement for provided services.⁷ As a result, there may be additional reimbursement options related to the provision of primary care services, and specifically ECM services.

While the data provided in this report are unprecedented, it is important to consider several project limitations when interpreting findings. First, there is some variation across project sites in providing and documenting services, particularly ECM services. To account for this limitation, we developed an algorithm that classified ECM services into five categories using NDW data on clinic, provider, and procedure codes; information from the project sites; and expert opinion. Despite use of this complex algorithm, ECM visits may be underreported. However, the ECM project findings may be used with additional information on provided services to further refine ECM definitions, and training may be conducted on the provision and documentation of such services.

A second limitation concerns data availability. CAIANH obtained the NDW extracts during month 16 of a project funded for 24 months. Thus, project timeframes limited the number and complexity of analyses. In addition, the IHS NDW did not include data on some services (e.g., pharmacy, laboratory) for a few project sites. Although efforts were made to obtain the data directly from the project sites, time constraints limited our ability to finalize these efforts during the project timeframe, and we were not able to report on some services for all 14 sites.

Third, CHS data include health service utilization information for services obtained at non-IHS providers yet paid for by IHS. Project findings do not include data on other non-IHS service utilization (e.g., number of visits, diagnosed conditions) and associated costs. Fourth, it is important to consider the context in which care is provided. For example, I/T providers may be more likely to admit a person to the hospital for observation, or recommend a longer length of hospital stay, due to geographic and economic considerations (e.g., distance to care, transportation resources) than non-IHS providers. Thus, evidenced-based guidelines (e.g., those related to hospital admissions) developed for other populations may require adaptation for use within IHS to account for the needs of the population it serves. Finally, the infrastructure includes a purposeful sample of AI/ANs identified as active users who lived in one of the 14 project sites, which were selected by specific criteria. While this approach provided an opportunity to include project site health system characteristics in the infrastructure, the population sample was not drawn randomly.

Mortality associated with diabetes and heart disease is one of many reasons that few AI/ANs survive into their seventies.^2,8-12 Mortality due to diabetes is 3-4 times higher among AI/ANs than among other Americans.^13-15 Although diabetes is the 7th leading cause of death in the U.S.,¹⁶ it ranks 4th among AI/ANs.¹⁷ The American Diabetes Association estimated that 40.1 percent of U.S. adults with diabetes were aged 65 years and older in 2012, and 27.5 percent were aged 70 years and older.¹⁸ According to FY2010 IHS Data Project findings, 26.0 percent of AI/AN adults with diabetes were aged 65 years and older; less than 17 percent were aged 70 years and older. Furthermore, CVD appears to be fatal more often in AI/ANs than in other populations.^8-10 AI/ANs have the highest rate of premature deaths from heart disease among all races.⁹ The AI/AN rate is nearly 2.5 times that for Whites, with 36.0 percent of deaths from heart disease occurring among persons aged younger than 65 years.⁹

IHS has implemented a number of initiatives, such as those implemented through the IPC Program and SDPI and under the guidance of the Pharmacy Program, to meet the needs of AI/ANs with chronic disease. It is hoped that these project findings may be used to identify additional opportunities to enhance the provision of services to improve AI/AN health outcomes.

^aDue to the length of the project’s contract title, IHS assigned this working project title to facilitate communication.
^bThe ��Ƶcontract was awarded to Denver Health. Denver Health subcontracted with CAIANH to implement the project and provided administrative project support.
^cIHS active users are persons who used IHS services at least once during the past three fiscal years.
^dNon-IHS providers are providers that are not I/T providers.
^eThe Health Information Committee included representation from each project site.
^fDiabetes education visits are visits that occurred in diabetes education clinics and are not counted in categories 2-4. The majority of these visits were conducted by nurses and health educators.
^gIt is important to note that, during one day, a person may have more than one clinic visit. For example, a person may have a general office, a dental, and an ECM visit.

References

¹U.S. Department of Health and Human Services, Indian Health Service. Improving patient care, improving the quality of and access to care. 2013; .

²U.S. Department of Health and Human Services, Indian Health Service. IHS Special Diabetes Program for Indians | 2011 Report to Congress: Making progress toward a healthier future. Rockville, MD. 2011.

³Giberson S, Yoder S, Lee MP. Improving patient and health system outcomes through advanced pharmacy practice. A report to the U.S. Surgeon General. Rockville, MD: U.S. Public Health Service, Office of the Chief Pharmacist; 2011.

⁴U.S. Department of Health and Human Services, Indian Health Service. Updated user population estimates FY 2010. Rockville, MD: 2011.

⁵O'Connell J, Yi R, Wilson C, et al. Racial disparities in health status: A comparison of the morbidity among American Indian and U.S. adults with diabetes. Diabetes Care. 2010;33:1-8.

⁶O'Connell JM, Wilson C, Manson SM, et al. The costs of treating American Indian adults with diabetes within the Indian Health Service. Am. J. Public Health. 2012;102(2):301-308.

⁷Rosenbaum SJ. The Patient Protection and Affordable Care Act: Implications for public health policy and practice. Public Health Rep. 2011;126(January-February):6.

⁸Rhoades DA. Racial misclassification and disparities in cardiovascular disease among American Indians and Alaska Natives. Circulation. 2005;111(10):1250-1256.

⁹Oh SS, Croft JB, Greenlund KJ, et al. Disparities in premature deaths from heart disease - 50 states and the District of Columbia, 2001. MMWR. 2004;53(6):121-125.

¹⁰Howard BV, Lee ET, Cowan LD, et al. Rising tide of cardiovascular disease in American Indians. The Strong Heart Study. Circulation. 1999;99(18):2389-2395.

¹¹Acton KJ, Burrows NR, Geiss LS, et al. Diabetes prevalence among American Indians and Alaska Natives and the overall population - United States, 1994-2002. MMWR. 2003;52(30):702-704.

¹²Crouch J, Korenbrot J, Korenbrot C. Medicare statistics for American Indians and Alaska Natives. Sacramento, CA: California Rural Indian Health Board; 2012.

¹³U.S. Department of Health and Human Services, Indian Health Service. Trends in Indian Health, 2000-2001. Rockville, MD: 2004.

¹⁴Grossman DC, Krieger JW, Sugarman JR, et al. Health status of urban American Indians and Alaska Natives. A population-based study. J. Am. Med. Assoc. 1994;271:845-850.

¹⁵Gilliland FD, Owen C, Gilliland SS, et al. Temporal trends in diabetes mortality among American Indians and Hispanics in New Mexico: Birth cohort and period effects. Am. J. Epidemiol. 1997;145:422-431.

¹⁶Hoyert DL, Xu JQ. Deaths: Preliminary data for 2011. Hyattsville, MD: National Center for Health Statistics; 2012.

¹⁷Heron M. Deaths: Leading causes for 2009. Hyattsville, MD: National Center for Health Statistics;2012.

¹⁸American Diabetes Association. Economic costs of diabetes in the U.S. in 2012. Diabetes Care. 2013;36:1033-1046.

SHARE:

ARRA ACTION: Comparative Effectiveness of Health Care Delivery Systems for American Indians and Alaska Natives Using Enhanced Data Infrastructure