In this sample of community-dwelling older adults, the SEM was applied to examine individual, interpersonal, institutional, and community factors related to hypertension awareness and control in Brazil and Colombia. While individual-level factors such as diabetes were associated with both hypertension awareness and control, associations were also observed with various institutional- and community-level factors. A notable finding was the significant country differences in factors associated with awareness and control. In Colombia, engagement in community activities was associated with higher hypertension awareness. The opposite was observed in Brazil. Similarly, there was no difference in hypertension awareness by age group in Colombia, while in Brazil, the older group (70–75 years) was more poorly controlled than the 64–69 group. This underscores the critical importance of studies within unique contexts27, as others conducting cross-cultural work have observed surprising differences in predictors of health outcomes across varying middle-income contexts.42,43
Hypertension was prevalent in both country samples, with approximately three quarters of Brazilian and Colombian IMIAS respondents either reporting a diagnosis or having a mean blood pressure of ≥ 140/90, taken as part of the study procedures. Prevalence estimates in this study were comparable to those of other low- and middle-income countries obtained as part of the WHO’s Study of Global Ageing and Adult Health (SAGE). In this study of over 35,000 adults 50 years and older from China, Ghana, India, Mexico, the Russian Federation and South Africa, 53% (range 32% in India to 78% in South Africa) had hypertension based on clinical measurement and self-report.43 Similarly, in a national study of over 20,000 older Colombians, 58% (95%CI 55 − 50) had hypertension.44 Because hypertension is a modifiable risk factor for most common causes of morbidity and mortality in older adults, these results highlight a need to improve the detection, treatment and control of hypertension among older adults in the LAC region.2,10,45
Among respondents with hypertension, awareness was high (> 80%) in both countries; awareness was slightly higher among Brazilian than Colombian respondents. This finding contrasts with the SAGE study. With the exception of the Russian Federation (72%), no more than 45% of respondents were aware of their hypertension in the other five countries.43 The WHO Region of the Americas, which includes Brazil and Colombia, has ambitious goals for the control of hypertension, including achieving an awareness level of 70% or more.2 Hypertension awareness in this study was higher than the 70% target. It was also higher than levels reported in the Cardiovascular Risk Factor Multiple Evaluation in Latin America (CARMELA) trial and the Perspective Urban Rural Epidemiological (PURE) study for South America.2 In CARMELA, which included respondents younger than the samples in this study, 31% of Bogotá respondents were unaware of their hypertension and a similar proportion of those with hypertension were controlled.45 Blood pressure control among respondents aware of their condition in this study was low, especially in Brazil, where less than a third of respondents achieved control. Similar to this study and CARMELA45, SAGE also reported very low levels of control among those aware of their hypertension.43
Hypertension awareness was strongly associated with use of health services, specifically frequent medical visits and taking medication. The high awareness observed among Brazilian and Colombian IMIAS respondents is consistent with the elevated levels of access to health services reported by them. In Colombia, 99% of respondents reported access to a usual source of care. In fact, because of the ubiquity of access to care in Colombia, this variable had to be excluded from the multivariate models. In Brazil, about two-thirds of respondents reported access to a usual source of care. Individuals who use health care services, especially those taking a medication or frequently visiting a provider, are doing so for a reason. Thus, strong associations between measures of health care utilization and hypertension awareness are not surprising, especially after adjusting for factors associated with health care utilization patterns such as sex, education and income. Others working with LAC populations have likewise reported strong associations between accessing care and hypertension awareness.10
There was a strong independent association observed between diabetes and hypertension. Hypertension and diabetes commonly occur together.46 Those who use health services regularly, especially if they have a comorbid condition like diabetes or obesity, are likely to get diagnosed with the comorbidity of hypertension. This would explain the very high proportion (> 95%) of individuals with diabetes who were aware of their hypertension status. Similar findings have been reported by others working in the LAC region.10
In a cross-sectional study of hypertension correlates, it is expected individuals with chronic conditions and those using health services would be aware of their hypertension status. Unexpectedly, there was an interaction between study site and community engagement on awareness. Brazilian respondents reporting engagement in community activities were less likely to be aware of their condition than those who did not. The opposite was observed in Colombia. One explanation is those who are busy engaging in community activities appear and feel healthy. Thus, they may consult with health care providers less frequently or providers perceive them as more robust than they actually are. While possible, it does not explain the site-specificity of the finding, which highlights the importance of considering contextual differences between communities. Previous research reports topics discussed within social networks influence diagnosis and control of hypertension. Individuals may benefit from one another’s advice and experiences regarding disease management, but if the group is unlikely to communicate about health, the risk for undiagnosed hypertension actually increases.39 The site-specificity of this study’s findings may reflect the nature and content of the community groups frequented at each site, including common topics of conversation.
This study also examined hypertension control and its correlates. Low levels of hypertension control have consistently been reported in Latin America.2,10,47,48 Notably, control was much lower in Brazil than Colombia, which is consistent with fewer Brazilian respondents reporting a usual source of care. This observation is corroborated by findings from a 2013 national survey of over 60,000 Brazilian adults which reported a significant association between access to care with both awareness and control. Critically, this large study also highlighted variations in the content of care provided to those with diagnosed hypertension, as well as uneven quality.10 Continuity of care is important for hypertension control; insufficient and/or inconsistent access to health services may critically interrupt treatment and management protocols, thereby resulting in poor control among those aware of their hypertension.
Women were significantly more likely than men to have controlled hypertension. Multiple studies report better hypertension control among women than men.7,9,10,43 This result may be mediated by better medication adherence, as women tend to be more adherent than men in some contexts,49,50 but not all51. Women also tend to be more proactive about accessing and using health services.52
Similar to the results for hypertension awareness, a community-level variable was associated with control—strolling shops and stores. Other research highlights the importance of social context to disease management and recognizes that much of the day-to-day work required to control hypertension occurs outside of the health care sector.39 Strolling shops and stores may capture exercise the respondents did not consider in this study’s walking tool or may represent an enjoyable, blood-pressure lowering activity enjoyed by older Latin American adults. It may also capture facets of community engagement and social resources that other variables did not. Irrespective, it is an intriguing finding consistent with other research highlighting the importance of social and community resources for hypertension management.39
A critical determinant of hypertension control is medication adherence among patients. The vast majority (> 85%) of study respondents reported taking medication in the past two weeks, 70% of whom were taking antihypertensives. And yet, blood pressure control was low. This may point to patient adherence issues. Despite the effectiveness of antihypertensive medications, adherence is a known challenge and half of the patients prescribed an antihypertensive drug stops taking it within one year.53 Low control may also suggest health system and provider deficiencies, such as failure to apply evidence-based clinical guidelines or prescription of less effective drugs.53 Given generally high rates of medication usage and high access to care (~ 100% in Colombia and 64% in Brazil), patient adherence issues alone are unlikely to explain the low overall control levels reported in this study and may point to a need to standardize treatment protocols, ensure access to affordable and effective drugs, and improve service delivery.2 These suggestions are reinforced by the observation that control was only marginally better among those with regular visits with their health care provider (although, given the cross-sectional study design, many of those with poor hypertension control may need to see their provider more often).
This study has many strengths, including the socioecological approach to factors other than demographic traits and behaviors, permitting exploration of how social context shapes diagnosis and management. It also has some limitations. First, the sample size in this study is modest and the numbers of individuals with certain outcomes (e.g. lack of hypertension awareness) were small. The sample size also limited our ability to examine potentially interesting and informative interactions such as between social support variables and sex, by country. Such analyses could be of interest in future work from better powered studies. Second, while the purpose of this study was solely to highlight correlates of hypertension awareness and control, it is critical to reiterate its cross-sectional nature and that our purpose is not to draw conclusions about cause and effect. Third, while self-reported measures of disease status are commonly used and validated, misclassification is still possible. This misclassification would most likely be non-differential and thus bias effect estimates towards the null. Finally, our sample contained older adults from two urban centers and while results can be generalized to the older adults of those cities, it may not be appropriate to generalize them further.