UC San Diego
Socio-ecological correlates of non-adherence to prescribed medications among low-income Hispanic/Latino adults with chronic conditions
- Author(s): Garcia, Melawhy Lucette
- Advisor(s): Talavera, Gregory A
- et al.
Background: Hispanic/Latinos experience high rates of uncontrolled chronic conditions that place this large segment of the population at risk for cardiovascular disease and mortality. Despite evidence that poor medication adherence is a major contributor to disparities in glycemic and lipid control; there is a gap in knowledge regarding the measurement and correlates of medication adherence among Hispanic/Latinos.
Aims: Using the World Health Organization’s Multidimensional Adherence Model, this dissertation focused on: 1) examining patient-, social/economic-, health condition-, therapy-, and health system-related correlates of poor adherence; and 2) identifying ways to effectively measure medication adherence among Hispanic/Latinos adults.
Methods: Aims 1 and 2 involved cross-sectional analyses of data on 149 participants enrolled in a randomized trial (Chapter 2); cross-sectional analyses of data on 279 participants enrolled in a randomized trial (Chapter 3); and, a historical cohort study of 1,485 patients from a federally qualified health center (Chapter 4).
Results: As measured by self-report, objective, and health record data, Hispanic/Latinos exhibit high rates of poor medication adherence. Bivariate analyses revealed numerous factors associated with poor adherence. Patient-related factors included younger age, sex, forgetting, multiple medications, and side effects. Health condition-related factors included higher depressive symptomatology, anxiety, and perceived stress. Social/economic-related factors included social support and limited health literacy. In logistic regression analyses, age was inversely related to poor adherence (OR 0.97; CI 0.96- 0.99, p < 0.01). Social support predicted low medication adherence (OR = 2.22; 1.03-4.76, p < 0.05) among males. The odds of low adherence were higher among males with limited health literacy (OR = 6.26; 1.06-37.10, p < 0.05). Comparison of measurement methods confirmed patients overestimate adherence to medications by self-report. Objective measures of adherence, including Medication Event Monitoring systems and Proportion of Days covered, showed lower adherence than self-report.
Conclusion: This study provides insight to the suboptimal levels of medication adherence among Hispanic/Latinos receiving services at federally qualified health centers. The findings from this study may help to inform future intervention research aiming to improve medication adherence among Hispanic/Latinos in a similar settings. Further research is needed to examine the multidimensional factors related to poor adherence among this population.