BACKGROUND/RATIONALE:
Ethnic minority populations bear an inordinate burden from type 2 diabetes as reflected by the higher occurrence of disease sequelae such as amputations, loss of vision, and premature mortality. Effective control of diabetes requires treatment with multiple agents and adherence to these medications is critical to achieve target treatment goals. Evidence supports that up to 50% of people with diabetes are non-adherent to medications. Medication non-adherence (MNA) is highly correlated with poor metabolic control, increased health care cost, and increased mortality. Prior to our study, there were significant gaps in knowledge on the relationship between MNA, race/ethnicity and subsequent health care costs and mortality in veterans with type 2 diabetes.

OBJECTIVE(S):
Three aims addressed the primary study objectives, which determined: 1) racial/ethnic differences in MNA in elderly veterans with type 2 diabetes after controlling for relevant covariates, 2) the effect of MNA on healthcare costs across racial/ethnic groups over time, and 3) the effect of MNA on mortality across racial/ethnic groups over time. We hypothesized that, compared to non-Hispanic Whites (NHW), ethnic minorities (non-Hispanic Black-NHB, and Hispanic) would have higher MNA, and higher healthcare costs and mortality with increased MNA over time.

METHODS:
Study Population: A national cohort of veterans with type 2 diabetes was created by linking multiple patient and administrative files. The National Patient Care (NPC) database was the source data for the VHA Medical SAS Datasets, used to analyze veteran clinical data such as diagnosis and procedure codes for inpatient and outpatient visits. The Pharmacy Benefits Management (PBM) database included utilization information for every prescription filled in the VA by a site's pharmacy (new fill or refill) and by a Consolidated Mail Outpatient Pharmacy. Veterans were included in the cohort if they had type 2 diabetes defined by two or more International Classification of Diseases, Ninth Revision (ICD-9) codes for diabetes (250, 357.2, 362.0, and 366.41) in the previous 24 months (2000 and 2001) or during 2002 from inpatient stays and/or outpatient visits on separate days. Veterans were followed from time of entry until death, loss to follow-up, or through December 2006. The cohort included a total of 892,223 veterans of which 61.51% were NHW, 12.14% were NHB, 13.86% were Hispanic and 12.48% had either missing or unknown race/ethnicity information.

Primary Variables of Interest: Annual medication possession ratio (MPR) was defined as the number of days supply divided by 365 days over the study period for each veteran. MPR was also calculated for individuals using 1) insulin combined with hypoglycemic agents, 2) insulin only, and 3) hypoglycemic agents only. Additionally, a three-level variable classified individuals as medication-adherent (i.e., MPR 80%), non-adherent and no medication. Finally, HbA1c was calculated as post-entry (baseline) mean of all observations recorded within a year. Regression models were used to model relationships between exposures and outcomes of interest.

FINDINGS/RESULTS:
Several key findings have emerged from the study. Three articles have been published in peer-reviewed journals and five manuscripts are currently under review. In article #1 titled "Regional, Geographic, and Ethnic Differences in Medication Adherence Among Adults with Type 2 Diabetes" we reported significant regional, rural/urban, and racial/ethnic differences in MPR and that rural/urban residence modified the effect of race/ethnicity on MPR. In article #2 titled "Regional, geographic, and racial/ethnic variation in glycemic control in a national sample of veterans with diabetes", we found that NHB and Hispanics had higher HbA1c levels compared to NHW, but differences were largely explained by adjustment for covariates. In article #3 titled "Using quantile regression to investigate racial disparities in medication non-adherence", we focused on using quantile regression to investigate racial disparities in MNA. The five manuscripts currently under review examine (1) the relationship between MPR and mortality stratified by racial/ethnic group; (2) the relationship between HbA1c levels and mortality stratified by medication adherence and racial/ethnic group; (3) methodological considerations when working with large databases; (4) the economic impact of MNA in veterans with diabetes; and (5) methodological advances in examining the economic impact of MNA. The collective evidence from these manuscripts confirm our hypotheses that veterans have a substantially higher mortality risk and healthcare costs in the lowest MPR quintile relative to the highest quintile. Additionally, we found evidence for racial/ethnic differences in the association between glycemic control and mortality, which varied by medication adherence. Joint modeling analyses found annual pharmacy cost increased approximately 3% per year (p=0.001) while inpatient and outpatient costs remained stable during the 5-year study time period. Compared to adherent subjects, patients with MNA had 43% higher annual inpatient costs, 6% lower outpatient costs, and 36% lower pharmacy costs.

IMPACT:
These findings suggest that racial/ethnic variations in MPR, glycemic control, and mortality exist nationally and that there are racial/ethnic differences in the ability of MPR and glycemic control to predict outcomes. For example, race/ethnicity and medication status (i.e., medication use and adherence) appear to modify the association between HbA1c and mortality and may be important to consider when setting individualized HbA1c targets. Moreover, our economic analysis provides information on the financial impact to the VA of MNA among veterans with type 2 diabetes. Based on our estimations, extrapolating to the full Veteran population with diabetes, achieving 90% medication adherence (defined as MPR>=0.9) would result in an estimated mean savings of $1.79 billion per year in 2011 dollars.

In summary, quality improvement initiatives should target improvements in medication adherence, identification and early initiation of appropriate diabetes medications, and aggressive management of comorbid conditions in Veterans with diabetes. The strong association between MNA and mortality warrants further attention to multi-level interventions that increase ongoing and consistent use of medications. Aggressive strategies and policies should be implemented to achieve optimal medication adherence in patients with Type 2 diabetes.

---

External Links for this Project

---

PUBLICATIONS:

None at this time.

DRA: Aging, Older Veterans' Health and Care, Diabetes and Other Endocrine Conditions
DRE: Treatment - Observational, Prognosis
Keywords: Adherence, Cost, Diabetes
MeSH Terms: none