Effectiveness of a provider and patient-focused intervention to improve hypertension management and control in the primary health care setting in Cuba: a controlled before-after study

About 17.8 million people die from cardiovascular diseases (CVDs) each year, 31% of all deaths worldwide [1], and more than 75% of those deaths occur in low- and middle-income countries (LMICs) [2]. Uncontrolled hypertension, the main modifiable risk factor for CVDs, is associated with over 10 million avoidable deaths annually [3]. Almost three-quarters of people with hypertension (650 million people) live in LMICs [4], where 88% of the hypertension-attributable mortality takes place [5]. In Latin America and the Caribbean, hypertension affects up to 40% of the adult population [6, 7], and CVDs annually lead to 1.8 million deaths in the region [8, 9].

The benefits of lowering blood pressure (BP) to prevent CVDs are well established [10‐14]. Notwithstanding, in a recent cross-sectional study on the management of hypertension in 44 LMICs [15], only 39.2% of individuals with hypertension were aware of their condition, 29.9% were being treated, and 10.3% had achieved BP control. A 2016 review including studies from 90 countries worldwide reported 26.3% hypertension control among patients aware of their condition in the represented LMICs, against 50.4% in the included high-income ones [16]. However, hypertension care and control figures show substantial variability among LMICs, corroborating the need to provide evidence on the contribution of risk factors, management strategies, and service delivery in their specific sociocultural and health care systems contexts [17]. Apart from the population’s possible low disease awareness and limited access to health care, most LMICs’ health systems are designed to provide acute curative care [18, 19]. Hence, chronic care is often reduced to fragmented and expensive disease management [20, 21], characterized by a lack of preventive approaches and oriented toward treating complications and acute exacerbations at the referral level [22, 23].

Yet, Primary Health Care (PHC) should play a key role in preventing and managing chronic conditions [22‐26], also by boosting the quality of the provider-patient interaction. Well-tuned provider-patient communication and shared decision-making have been associated with better adherence to anti-hypertensive treatment and improved self-care management in diverse countries and populations [27, 28], including low-income black hypertensive patients in the United States [29, 30], patients attending PHC services in Latin America [31, 32] and patients treated in specialist services in Europe [33]. In addition, also provider-patient communication that addresses patients’ sociodemographic circumstances or promotes social support is instrumental in enhancing medication adherence [34, 35].

International initiatives such as the Standardized Hypertension Treatment and Prevention Project [19], developed by the Pan American Health Organization and the United States Centers for Disease Control and Prevention, have identified six essential components for intervention to improve hypertension control: guideline-based standardized treatment protocols; access to essential medicines and technology; health information systems for cohort monitoring and evaluation; patient empowerment; team-based care; and community engagement. The Global Hearts Initiative [36], led by the World Health Organization, stresses the need to strengthen health systems at the primary care level. The Initiative’s technical package addresses the four main risk factors for CVDs (tobacco use, physical inactivity, unhealthy diets, and harmful alcohol consumption). It further emphasizes risk-based management, appropriate referral, improved service delivery by task-sharing, and robust clinical monitoring. Interventions along the above lines, directed at the community, patients, and health care providers, have undoubtedly the potential to improve outcomes for people with chronic conditions in general and hypertension in particular. Still, there is a lack of evidence on their effectiveness from research conducted in LMICs [37]. Studies in Mexico [38, 39], Argentina [40], and Peru [41] are noteworthy exceptions in Latin America and the Caribbean.

In Cuba, a middle-income country, chronic conditions and CVDs account for 84 and 36% of the overall mortality, respectively [42]. Hypertension prevalence, in turn, was estimated at 30.9% in the population above 15 years in the most recent nationwide survey [43]. In contrast to the majority of fragmented Latin American health systems, the Cuban national system provides integrated care through an accessible network of family practices. It is internationally recognized for being well-organized, based on a PHC approach, and enjoying favorable health indicators [44‐46]. In recent decades, the Cuban health system has shown remarkable hypertension control figures at the population level, similar to those reported by the United States and Canada [47]. It was signaled as an early example of significant success in controlling hypertension [48] and, at the same time, put forward as an appropriate setting to study the potential of resource-constrained health systems in improving hypertension treatment and control [49]. Indeed, the Cuban health system still faces significant challenges in comprehensively addressing non-communicable disease prevention and control. Despite free and universal health services, there are utilization and supply gaps, and there is scope for tweaking the quality of care [47, 50].

Our 2012 study amongst individuals with diagnosed hypertension in the municipalities of Cardenas and Santiago documented that 91% were on pharmacological treatment but only 58% had controlled hypertension [25]. Appraisal of the population’s representations regarding hypertension [51] documented unsatisfactory knowledge of hypertension management, self-care requirements, and potential complications. Audits of family doctors’ clinical records and assessment of the quality of the provided care [52] uncovered significant gaps in the PHC staff’s clinical performance and poor adherence to the national guidelines for hypertension management clinical inertia, and inadequate prescription of anti-hypertensive drugs. Based on this formative research, we set up an intervention to reduce the hypertension control gap, targeting primary care practitioners and hypertensive patients under their care. For the former, we focused on increasing compliance with evidence-based standardized guidelines, enhancing clinical skills, and improving the quality of the provider-patient interaction. For patients, the intervention primarily aimed to augment their awareness of the need for preventing complications, foster self-care skills, and improve hypertension treatment adherence. The present paper’s primary objective is to assess the effectiveness of the intervention in increasing the proportion of patients with controlled hypertension.

All hypertensive patients that were successfully contacted consented to participate in the surveys. However, in the baseline survey, sixty-seven sampled patients (5%), mainly from the Echeverria health area, were not encountered at home after two repeat visits and were not replaced due to operational constraints. Hence, the total number of respondents was 1333. The endline survey included 1404 respondents. We excluded 4 cases with a missing value for education or obesity, leaving 2733 participants in the analyses (Table 1).

The mean age (SD) of participants in the baseline and endline survey was 59.8 (14.0) and 60.9 (14.5) years, respectively. In both surveys, there were 1.6 times more women than men. Table 1 provides further details on the general characteristics of the included patients. Notably, most participants were on pharmacological anti-hypertensive treatment, but the percentage of non-adherence was high. In addition, the adherence increased substantially from baseline to endline in both intervention arms but not in the control arm.

Seventy-three family doctors from 61 intervention FD&N practices attended the provider workshops. This number includes staff present at the start and possible substitutes for family doctors posted elsewhere or participating in international missions who were trained in September 2013. Nurses were posted much more stably, but only 54 attended, mainly due to the workshops being organized outside working hours. Eventually, 3308 hypertensive patients in the ‘Provider and Patient’ study arm, 51.6% of the eligible ones, participated in the hypertension schools. The main reason for not attending was a conflict between the schedule of the sessions and patients’ labor obligations.

In Table 2, we present the crude intervention effect. In absolute terms, the difference in % endline minus % baseline hypertension control was 18.7% higher in the study’s intervention arms than in the control arm (p < 0.001). The odds of attaining hypertension control in the second survey round were more than twice higher (p < 0.001) in the intervention arms than in control.

After controlling for potential confounding factors in multivariable logistic regression (Table 3), the odds of having attained controlled hypertension remained twice or higher in patients in the study’s intervention arms than in the control arm (study arm by survey round interaction terms; p < 0.001). The point estimate of the intervention effect was somewhat higher for the provider only as compared to the provider and patient intervention arm, but confidence intervals largely overlapped. In both municipalities, the intervention components were allocated to the health areas with the lower baseline hypertension control prevalence, which is reflected by the ORs below 1 under the study arm variable. The OR for the endline survey round is not significantly different from 1, indicating no baseline to endline change in the control arm. Considering the whole study population, having post-primary education, not being obese, and being of white ethnicity were independently and positively associated –be it rather weakly- with hypertension control.

This controlled before-after study in 122 FD&N practices in two Cuban municipalities found that an intervention targeting PHC providers resulted in a substantially increased proportion of patients with controlled hypertension. However, adding to the provider component a behavioral intervention that targeted the hypertensive patients under their care did not augment the effect. In both intervention arms of the study, the adherence to anti-hypertensive pharmacological treatment improved markedly from baseline to endline, and the average patients’ systolic and diastolic BP decreased. After the intervention, the adjusted odds ratio of having controlled hypertension was twice higher in the study’s intervention arms than in the usual care arm.

Our findings can be compared with the results obtained in a one-year, uncontrolled HEARTS [59] intervention led by the Pan American Health Organization in the catchment area of a policlinic in the Matanzas province, Cuba [60]. It improved hypertension control in registered patients by almost 10%, from 59.3 to 68.4%. Averaging over both intervention arms, we attained close to a 20% difference of difference increase in the intervention arms of our controlled study. A multi-component cluster-randomized trial conducted in public primary care centers in Argentina [40] resulted, after 18 months, in a difference of difference of about 21% hypertension control increase in the intervention group compared to the usual care group. However, its baseline control prevalence was only 17% against over 50% in our study.

In a recent non-controlled patient-centered intervention conducted in three health centers in Lima, Peru [41], the use of BP lowering medication increased from 73 to 82% after 24 months. In our intervention, the vast majority of patients were already on pharmacological treatment at baseline, but the proportion with adequate adherence increased by around 20%. We hypothesize that the training of the FD&N practitioners on recommended clinical and treatment guidelines [61‐63] enhanced prescribing of standardized anti-hypertensive treatment regimens, as was observed in a trial conducted in Nigeria and China [64]. At the same time, it should have improved their skills for interacting with and motivating patients, which could have contributed to boosting patients’ medication adherence.

A review of health care provider-led interventions on lifestyle modification in hypertensive patients in high-income countries [65] supports the idea that this approach can contribute to lowering elevated BP. Therefore, the apparent lack of increased hypertension control when adding the intervention component targeting hypertensive patients on top of our provider component is somewhat surprising. However, only around 52% of patients attended hypertension schools. Additionally, the serial small group organization of the schools delayed building up coverage. In contrast, in Costa Rica, a country similar to Cuba in terms of PHC approach and structure, a healthy lifestyle intervention to reduce the cardiovascular disease risk of hypertensive and diabetic patients resulted, after adjusting for confounding variables, in a significant average 8 and 5 mmHg reduction in systolic and diastolic BP, respectively, despite only 58% of the patients enrolled in the intervention group having attended at least one group education session [66].

A possible explanation for the lack of effect of our educational component is that we targeted patients with uncomplicated hypertension, possibly excluding the ones which might have benefitted more: the individuals at higher CVDs risk that require intensive and frequent lifestyle counseling [36], who were already under specialist care. On the other hand, in our hypertension schools, we relied on interactive educational group sessions, which are more effective than self-learning by reading and classic lectures as educational strategies for attaining behavioral change [67]. However, on the pathway between exposure to the intervention and hypertension control, lifestyle modifications are subject to underlying and interrelated behavioral, nutritional, social, and environmental drivers, and change remains uncertain [5].

Unsurprisingly, a Cochrane systematic review [53] found that isolated educational interventions unlikely lead to significant BP reductions. However, it also indicated that nurse-led care might have potential and that successful hypertension control trials were always multi-component and included, as critical elements, registration and regular follow-up of patients and a stepped approach to anti-hypertensive treatment. A more recent review [68] corroborates that multi-component strategies are the most effective, alongside team-based care, but notes that data from LMICs are sparse. Finally, it has also been pointed out that task sharing or shifting with non-physician health care workers in these countries may produce favorable effects on BP control [69].

Our study has some limitations. First, given the integrated nature of the Cuban health system and close social networks within the society at large, some cross-contamination between study arms might have occurred. However, the control (usual care) and intervention activities were allocated to administrative and operationally independent health areas and BWGs, which put up a barrier against it. Of note, contamination -if any- will have diluted the intervention effect. Second, the output of the different elements making up the intervention components was not measured separately, and we cannot sort out which ones may have contributed most to the outcome. Third, the inclusion of process evaluations has been encouraged for assessing multi-component interventions [37]. We controlled the delivery of the components of our intervention regarding content and targeted population but did not collect structured information on the process and fidelity of their implementation. This may hamper the interpretation of our results and compromise the ease of scaling up the intervention. Finally, from a subject matter point of view, the provider component somewhat neglected the training on cardiovascular risk assessment and the use of treatment algorithms.

The study also has clear strengths. First, it is among a few studies in LMICs [38‐41, 70] that evaluate the effectiveness of a multi-component intervention to improve hypertension control by building on the potential of PHC. The Cuban public health system is homogeneous and provides universal access and integrated care through a dense nationwide PHC network. Our results should be replicable nationally but the intervention may not be transferable to other LMICs without adapting its components and activities to the local context and prevailing bottlenecks for improving hypertension control. On the other hand, given the poor control levels currently attained in many resource-constraints settings, the opportunities to pick low-hanging fruit may be considerable. Second, it was preceded by formative research [25, 51, 52] and developed with the active participation of the Hypertension Program of the Cuban Ministry of Health, local health system staff, and the communities belonging to the catchment areas of the involved FD&N practices -in line with recommendations to incorporate local communities as co-designers of interventions tailored to their needs [71]. Third, the intervention was carried out by existing PHC staff, minimizing the need to allocate additional resources and enhancing the potential for scaling up.

We demonstrate that taking inspiration from the World Health Organization’s Hearts Initiative, it is feasible to substantially improve hypertension outcomes by intervention at the PHC level, even in a health care system that has already achieved a relatively high control rate. In preparation for scaling up the intervention to other health areas, further implementation research should tackle the challenges ahead: strengthening cardiovascular risk assessment and the use of simple treatment algorithms in the provider component, fine-tuning the behavioral change messages in the patient component, complementing the intervention with a community component targeting the public at large, and studying in-depth the implementation process.