The second goal of this project was to conduct a CER study to assess strategies implemented to reduce CVD risk among AI/ANs with diabetes and/or CVD. 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, analyses were conducted to document the types of ECM services provided and examine patient characteristics associated with their use by adults with diabetes or CVD. These 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.m
Tables 23-26 include information on utilization of ECM services by health status and project site for FY2010. The first of these tables, Table 23, provides data for all project sites by health status. The average number of ECM visits by adults with diabetes was 1.2. Among these adults, 41.1 percent had one or more ECM visits; their average number of ECM visits was 2.8. Table 23, also provides data on the percentage of persons who had one, two, three, or four or more visits. Among adults with diabetes who had at least one visit, 42.3 percent had one visit, 21.3 percent had two visits, 13.9 percent had three visits, and 22.5 percent had four or more visits. Adults with both diabetes and CVD were more likely to use ECM services and had a higher number of visits. While only 10.6 percent of adults with CVD but not diabetes used ECM services, the average number of visits among those who did was 4.5.
Table 24 provides similar information for adults with diabetes by project site. Although the percentage of adults with diabetes who had at least one ECM visit averaged 41.1 percent for all project sites, there was variation in this measure across the 14 project sites, with values ranging from under 30 percent to over 70 percent. Reasons for this variation included differences in the provision of ECM services (e.g., IPC model differences, Tribal health program provided services in a Service Unit operated by IHS) and documentation of services. Among adults with diabetes who had at least one visit, the average number of ECM visits also varied by project site. The average number of visits ranged from under two to four or more.
Information on the use of ECM services by type of visit is provided in Table 25 for all project sites. Among adults with diabetes who had at least one visit, the average number of visits was 2.8. The average number of diabetes education, nutrition education, clinical pharmacy, case management, and other health education visits was 0.5, 0.9, 1.0, 0.1, and 0.3, respectively.
Among adults with both diabetes and CVD, ECM utilization was similar to all adults with diabetes, except that those with both conditions had higher use of clinical pharmacy services. Among those with CVD but not diabetes, the average number of visits was 4.5 among those who had at least one visit. The majority of these visits were for clinical pharmacy services.
Table 26 provides site-specific ECM findings for adults with diabetes. In some sites the majority of ECM visits were classified as diabetes education, while in other sites the majority of visits were provided by dieticians/nutritionists or clinical pharmacists. 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, although this percentage also varied by project site.
A second task related to this goal was to assess characteristics associated with ECM utilization by adults with diabetes or CVD. See Table 27. Among all adults with diabetes or CVD, the average number of ECM visits during FY2010 was 1.0. Nearly one-third used ECM services; the average number of ECM visits among users was 3.0.
In general, ECM use increased with age. Among adults aged 44 years or younger, a smaller percentage used ECM than among those aged 55 to 64 years. Differences in ECM use between males and females were not observed. ECM utilization also varied by type of health coverage. A larger percentage of adults with Medicare or private coverage used ECM services than those with Medicaid or no coverage.
As noted above, ECM use varied by health status. Only 10.6 percent of adults with CVD but not diabetes used ECM services during FY2010. The percentage of adults with diabetes and CVD who used ECM services was 44.2 percent. It is important to consider that many adults with diabetes or CVD had other chronic conditions. For example, the prevalence of hypertension among adults with diabetes but not CVD, diabetes and CVD, and CVD but not diabetes was 48.8 percent, 90.9 percent, and 62.2 percent, respectively. Over 36 percent of those with hypertension used ECM services
Among adults with diabetes or CVD, 24.3 percent had a mental health disorder and 6.0 percent had a substance use disorder. The prevalence of ECM use among adults with diabetes or CVD and a mental health disorder was 36.4 percent. Among the adults with diabetes or CVD and a substance use disorder, 25.6 percent had one or more ECM visits.
Health risk scores were categorized into quartiles, with 25 percent of all adult AI/ANs assigned to each of the following four categories: low risk, moderate risk, moderately high risk, and high risk. ECM use increased as health risk increased among adults with diabetes or CVD. Ten percent or fewer adults with low or moderate risk used ECM services. Over 30 percent of those at moderately high or high risk did. Furthermore, the average number of visits was lower among those at low or moderate risk as compared to those at moderately high or high risk.
A multivariate logistic regression was estimated to examine the statistical influence of patient characteristics and organizational type on the likelihood of using ECM services (i.e., having one or more visits). The logistic model was estimated using robust standard errors to account for correlation among AI/ANs within the same project site. We did not include a measure of health risk in the logistic model, as health risk is based on age, gender, and health status and specific measures of each were included in the model. When the health risk measures were added to the model, they did not statistically increase the ability of the model to explain variation in use of ECM services.
Findings from the logistic regression are presented in Table 28. Controlling for health coverage, health status, and other factors, the differences in ECM use by age that were noted above were not found to be statistically significant. Adults with private coverage were 20 percent more likely to have an ECM visit than those with no coverage. Differences in the likelihood of using ECM services between those with Medicaid coverage and no coverage were not found to be statistically significant.
Adults with CVD but not diabetes were significantly less likely to have an ECM visit than adults with diabetes but not CVD, similar to the descriptive findings. No differences were observed between adults with diabetes and adults with both diabetes and CVD. Controlling for other factors, including the presence of diabetes and CVD, adults with hypertension and adults with a mental health disorder were more likely to use ECM services than those who did not have these conditions. Adults with hypertension were 57 percent more likely to use ECM services than those who did not have hypertension. Adults with a mental health disorder were 20 percent more likely to do so.
The influence of two additional factors on ECM use was assessed. While organization type was not found to influence the likelihood of using ECM services, use of primary/general office and diabetes clinic services positively influenced use. In other words, the findings indicate that adults who are less likely to use primary/general office and diabetes clinic services are less likely to use ECM services.
Information on patient characteristics associated with use of ECM services may be used to understand who is accessing services and their level of service use (i.e., the number of visits). It 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. Finally, the ECM findings presented in this chapter will inform CAIANH’s future work on the CER study. CAIANH will examine the influence of ECM utilization, by provider type, on patient treatment costs for the CER study, using statistical models that control for issues related to selection bias in observational studies.