Medication Adherence Is a Mediator of the Relationship between Ethnicity and Event-Free Survival in Patients with Heart Failure

Introduction

Heart failure (HF) has been described as a “new epidemic”. The number of people diagnosed with HF is expected to double within the next 25 years owing to the aging of the population and improvements in treatment for cardiac conditions resulting in increased survival of cardiovascular diseases.^{1, 2} Acute exacerbation of HF signs and symptoms commonly requires hospitalizations that may contribute to increased mortality.³

The natural history of HF can now be modified by appropriate pharmacological therapy.^4–6 However, pharmacological therapy only benefits patients who take their medications as prescribed. Therefore, medication adherence is essential to achieve better health outcomes.^7–14 Poor medication adherence increases the risk of mortality and morbidity¹⁵ and leads to high health care costs^{8, 12, 15} in patients with HF.

There is a higher prevalence of HF among African-Americans compared with the general population (3% vs 2%, respectively).^{16, 17} African-Americans tend to be diagnosed with HF at an earlier age, have lower left ventricular ejection fractions (LVEF), and worse New York Heart Association (NYHA) functional class at the time of diagnosis than Caucasians.^{16, 18–20} Moreover, African-American HF patients have higher rehospitalization rates,^21–26 longer lengths of hospital stay²⁵ and higher hospital charges²⁵ than Caucasian patients. The reasons for the disparity in outcomes between African-Americans and Caucasians are unclear. Prior researchers have suggested that African-American patients with HF have lower adherence rates than Caucasian patients with HF^{27, 28} but whether differences in medication adherence may play a role in disparity in outcomes is unknown. Accordingly, the purpose of this study was to determine whether medication adherence is a mediator of the relationship between ethnicity and the composite endpoint of event-free survival of ED visits for HF exacerbation or cardiac or HF hospitalization or mortality in patients with HF.

Methods

Results

Discussion

A goal of Healthy People 2010 is the elimination of racial/ethnic health disparities.⁵² African-American patients with HF have a higher rate of hospital readmission than Caucasian patients.^{22–24, 26} This finding suggests that additional strategies that address factors underlying this disparity in African-American HF patients may be needed. It is important to explore factors associated with racial/ethnic health disparities in patients with HF so that effective interventions to health disparities can be developed. To our knowledge this is the first study to examine mediators of the link between ethnicity and outcomes in HF patients. Our results demonstrated that medication adherence was a mediator of the relationship between ethnicity and event-free survival in patients with HF.

Consistent with other studies, we found that the rates of HF or cardiac rehospitalization were higher in African-American patients than Caucasian patients.^{21–26, 53, 54} However, reasons for the increased risk for hospitalization in African-Americans compared to Caucasians are unclear. Biological differences between the two ethnic groups and/or differences in psychosocial and behavioral factors may play a role on racial/ethnic disparity in outcomes.^{24, 26, 55, 56}

Currently, there is no direct evidence that biological differences among ethnicities contribute to disparity in outcomes. In some studies, investigators found genetic differences that contribute to the development of HF and identified differences in the etiology of HF between African-Americans and Caucasians. These findings provide some evidence to explain racial/ethnic disparity in outcomes. Advocates of a biological explanation for disparities point to advances in genomic techniques, that allowed investigation of genetic factors that might contribute to racial/ethnic disparity. In one case-comparison study,⁵⁵ African-American participants with variant alpha_2c receptor genotypes were 5 times more likely to have HF than African-Americans with other receptor genotypes. No HF risk was associated with the variant beta₁ receptor alone; but, participants who had both variant receptor genotypes had a tenfold increased risk of developing HF. However, to date, there is no direct genetic explanation for the disparity in hospitalization rates.

Another biological explanation is the etiology of HF. Hypertension is the most common cause of HF in African American HF patients and coronary heart disease (CHD) is the most common cause of HF in Caucasian HF patients.^{16, 17, 19, 57–59} In the large Studies of Left Ventricular Dysfunction (SOLVD) registry, hypertension was found as the etiology of HF in 32% of African Americans and only 4% of Caucasians, while CHD was the etiology in 36% of African-Americans and 73% of Caucasians.⁵⁶ In our study, hypertension was the cause of HF in 18% of African-Americans and only 5.5% of Caucasians, while CHD was the cause of HF in 35% of African-Americans and 63% of Caucasians. In a large study,⁶⁰ investigators examined the natural history of HF with preserved left ventricular systolic function in African-American and Caucasian patients over five years. African American patients had a significantly higher mortality risk than white patients (hazard ratio [HR] = 1.34). The racial difference in survival rate was most prominent in patients with a non-ischemic etiology (HR = 1.6) compared with patients with ischemic heart failure. The finding that HF etiology is different between these two ethnic groups and non-ischemic etiology of HF leads to worse outcomes compared to ischemic etiology of HF provides other support for a biological explanation. However, without direct evidence, it is hard to conclude that racial/ethnic disparity in outcomes is due to biological differences. Therefore, more studies are needed to explore direct biological evidence of racial/ethnic difference in outcomes.

Other researchers have reported that differences in rehospitalizations may be related to difference in the prevalence of comorbid conditions and the effectiveness of their treatment,^{23, 24} differences in socioeconomic factors (e.g., age, access to care),^{23, 61} differences in psychological factors (e.g., depression),²⁴ and differences in adherence to diet and prescribed medication.²⁶ In our study, medication adherence emerged as a predominant mechanism linking ethnicity and poor outcomes. When we compared sociodemographic and clinical variables between African-American and Caucasian participants, only gender and medication adherence were different between these two ethnic groups. Although the African-American group had a greater percentage of female participants than Caucasian group, gender was not related to event-free survival. This result is consistent with prior studies showing no gender difference in rehospitalization or mortality in HF.^{62, 63}

A major finding in our study was that patients in the Caucasian group were more likely to be adherent than those in the African-American group. Studies of the relationship between ethnicity and medication adherence have produced inconsistent results.^{27, 32, 64–66} A number of investigators reported no differences in medication adherence based on race/ethnicity.^{32, 64, 67, 68} However, in three studies,^{27, 65, 66} African-American participants were less adherent than Caucasian participants. Each of these studies had one or more of the following limitations that may have influenced their results: self-report measures of medication adherence,^{64, 67, 68} small sample sizes,^{32, 64} or lack of consistent definitions of medication adherence.^{27, 32, 64–68}

It is unclear why African-Americans might have lower adherence. Investigators have postulated that it is not race alone, but the interaction of race and income that is related to adherence.⁶⁹ Other investigators found that African-Americans were more likely to experience side effects from taking medication compared to Caucasians.⁷⁰ Differences in adherence also may be related to differences in depression⁷¹ or out-of-pocket prescription costs.⁷² In a national study,⁷³ African-Americans were 1.38 times more likely to report cost-related nonadherence compared to Caucasians.

In prior studies, investigators reported attitudes toward medication,^{74, 75} knowledge,^75–77 and perceived barriers to medication taking^{78, 79} were correlated to medication adherence. African-Americans were more likely to have negative attitudes toward medication,⁸⁰ lower levels of knowledge about medications,^{81, 82} and more barriers to medication taking.⁸³ Low levels of health literacy⁸¹ may account for lower levels of health-related knowledge among African-Americans. However, in another study,⁸⁴ African-American patients reported that their health care providers were more active in advising and counseling about hypertension care and medication adherence than did Caucasians. African-Americans were found to have more knowledge of the risks/benefits of available therapies, to be more aware of the importance of controlling their blood pressure, and to be equally adherent. These data suggest that disparities in health-related knowledge and attitudes between African-Americans and Caucasians can be eliminated when both patients and health care providers are actively engaged in the health care process.

The association between medication nonadherence and ethnicity could be explained by income, side effects of medication, depression, out-of-pocket prescription costs, cost of medication, attitudes toward medication taking, knowledge of medication and barriers to medication adherence. Our results suggest the need to collect more data of potential factors related to racial/ethnic disparity on medication adherence to provide further insight into the relationships among race/ethnicity, medication adherence and outcomes and to improve medication adherence and reduce racial/ethnic disparity in outcomes.

Limitations

Although we demonstrated significant differences between African-Americans and Caucasians, a larger sample size of African-American HF patients is needed to generalize these results. Our sample only included 10% African-Americans and future studies of this phenomenon should include a larger proportion of African-Americans so that the complex dynamics surrounding adherence and outcomes can be better illuminated. The small sample of African-Americans also may produce unstable results that need to be verified in a larger sample. Thus our findings should be considered exploratory and the need for replication emphasized.

Conclusion

The major finding of this study was that medication adherence was a mediator of the relationship between ethnicity and event-free survival in this sample. Although these data suggest that interventions designed to improve medication adherence in African-Americans may decrease the disparity in outcomes between African-American and Caucasian patients with HF, our findings must be replicated before definitive recommendations can be made.