The 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 FY2010 project population included persons who lived in 14 project sites, 28.7 percent of all FY2010 IHS active users.9 This was the first time that such data were available for a large percentage of the population served by IHS. Previous to this project, IHS had such data for only one Service Unit.3, 10 Key FY2010 findings include:
- 14.7 percent of adults 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.
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 ECM services. Finally, the treatment cost findings will increase understanding of existing patterns of health resource allocation.
Results on third-party health coverage 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.11 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. Similarly, it may be possible to refine definitions and documentation of case management services provided telephonically, in order to more accurately assess the provision of such care.
A second limitation concerns data availability and quality. 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. Due to these constraints, we were not able to fully implement a longitudinal CER study to examine the influence of the provision of ECM services by different professionals (e.g., registered dieticians, pharmacists, certified diabetes educators) on treatment costs. In lieu of conducting a longitudinal CER study, analyses were conducted to examine patient characteristics associated with use of ECM services among adults with diabetes or CVD. The findings will inform future work on the CER study. 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. The quality of the diagnostic code data on the utilization records may have also varied, influencing health status measures based on the codes. Additionally, project site results for laboratory tests (e.g., LDL cholesterol) may not all have been stored in the same laboratory data field. Consequently, some laboratory data may not have been uploaded by each project site to the NDW, and laboratory findings may be underreported for some project 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. Because the project was designed to examine service use and treatment costs within IHS, we believe this limitation does not adversely impact the implications of this project.n
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. IHS active users are persons who had at least one outpatient visit during the past three years and, consequently, are more likely to have poorer health and to be female than AI/ANs living in the project sites who did not use services. Except in the California and Portland IHS Areas, the project sites had at least one inpatient facility, whereas some IHS Service Units do not provide inpatient services. Consequently, the estimated average treatment costs for the project sites may be higher than costs in Service Units without inpatient facilities.
Mortality associated with diabetes and heart disease is one of many reasons that few AI/ANs survive into their seventies.2,12-16 Mortality due to diabetes is 3-4 times higher among AI/ANs than among other Americans.17-19 Although diabetes is the 7th leading cause of death in the U.S.,20 it ranks 4th among AI/ANs.21 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.22 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.12-14 AI/ANs have the highest rate of premature deaths from heart disease among all races.13 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.13
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 and ensure effective use of I/T health program resources.
n In such situations, reports based on data compiled with Service Unit data, such as the IHS Diabetes Care and Outcomes Audit, rather than with NDW data, may be referenced.
