A Narrative Synthesis of the Health Systems Factors Influencing Optimal Hypertension Control in Sub-Saharan Africa

Introduction In sub-Saharan Africa (SSA), an estimated 74.7 million individuals are hypertensive. Reducing the growing burden of hypertension in sub-Saharan Africa will require a variety of strategies one of which is identifying the extent to which actions originating at the health systems level improves optimal management and control. Methods and Results We conducted a narrative synthesis of available papers examining health systems factors influencing optimal hypertension in SSA. Eligible studies included those that analyzed the impact of health systems on hypertension awareness, treatment, control and medication adherence. Twenty-five articles met the inclusion criteria and the narrative synthesis identified the following themes: 1) how physical resources influence mechanisms supportive of optimal hypertension control; 2) the role of human resources with enabling and/or inhibiting hypertension control goals; 3) the availability and/or use of intellectual resources; 4) how health systems financing facilitate and/or compromise access to products necessary for optimal hypertension control. Conclusion The findings highlight the need for further research on the health systems factors that influence management and control of hypertension in the region.


Introduction
In sub-Saharan Africa (SSA), an estimated 74.7 million individuals are hypertensive. Reducing the growing burden of hypertension in sub-Saharan Africa will require a variety of strategies one of which is identifying the extent to which actions originating at the health systems level improves optimal management and control.

Methods and Results
We conducted a narrative synthesis of available papers examining health systems factors influencing optimal hypertension in SSA. Eligible studies included those that analyzed the impact of health systems on hypertension awareness, treatment, control and medication adherence. Twenty-five articles met the inclusion criteria and the narrative synthesis identified the following themes: 1) how physical resources influence mechanisms supportive of optimal hypertension control; 2) the role of human resources with enabling and/or inhibiting hypertension control goals; 3) the availability and/or use of intellectual resources; 4) how health systems financing facilitate and/or compromise access to products necessary for optimal hypertension control.

Conclusion
The findings highlight the need for further research on the health systems factors that influence management and control of hypertension in the region.

Introduction
generalizations about health systems, including the perspective that all health systems can be "defined as a mere list of different organizations or persons that participate in producing health services." Such a definition, he argues, ignores for example, the population which is an essential part of the system [23].
Despite these shortcomings, there is a lack of health systems analysis as applied to hypertension in SSA using either the WHO building blocks or any other model. Instead, recommended actions to improve health systems in SSA often include the need to strengthen health financing [24] and/or removal of user fees [25,26]. However, it is unknown whether other factors and not only user fees or adequate financing alone, play a role with facilitating or hampering access to health systems services that are necessary for achieving optimal hypertension control [25,26]. Given the impact of hypertension, projected to become the most consequential disease burden by 2025 in SSA, the aim of this paper is to critically examine the health systems factors influencing hypertension awareness, treatment, control, and medication adherence in the region.

Methods Design
Using Popay and colleagues [27] as a guide, we conducted a narrative synthesis of available papers that examined the health systems factors influencing optimal management of hypertension in SSA. Narrative synthesis refers to 'an approach to the systematic review and synthesis of findings from multiple sources and relies primarily on the use of words and text to summarize and explain the findings of the synthesis' [27]. It is used when statistical meta-analysis or another specialist form of synthesis (such as meta-ethnography for qualitative studies) is not feasible particularly due to extreme heterogeneity in the methodological descriptions of available studies [27]. A narrative synthesis was appropriate for this review given an initial scoping exercise that revealed that the literature was too heterogeneous to permit a meta-analysis [4].

Data Sources
Searches of articles were conducted on the PUBMED database. The search was conducted for the period from January 2013 to September 2014. We used the search terms (see Table 1 and Appendix) "hypertension" and "Sub-Saharan Africa" or "" or "blood pressure" and "Africa" in combination with factors we considered as "elements of the health systems"" (See Fig 1 and S1 Research Protocol for more details). The review methodology was also adapted from other reviews focused on the influence of health systems on hypertension outcomes [4] which we defined as hypertension awareness, treatment, medication adherence, and control (the achievement of BP below 140/90mm Hg). We also retrieved articles from the reference lists of available studies.

Developing a narrative synthesis
We followed the steps delineated by Popay and colleagues [27], namely (1) developing (and/or) identifying a theoretical model, (2) developing a preliminary analysis, (3), exploration of relationships in the data, and (4), assessment of the robustness of the synthesis. Rather than developing a theoretical model as outlined Popay [27], we built on the schematic model developed by Maimaris et al. [4] which consists of four domains relating to key system level inputs that are required for effective management of chronic diseases such a hypertension (See Fig 1). Specifically, "the model aims to capture the complex interactions and inter-relationships that exist between the elements within a health system. It also highlights the important role that context plays in shaping the relationships between health system inputs and outcomes, while recognizing the complex adaptive nature of health systems [4]." Preliminary analysis of the available literature consisted of extracting the descriptive characteristics of retrieved articles in a table so as to produce a textual summary of the results. This enabled the exploration of relationships both within and between the studies reviewed as well as quality appraisal of the methodology used in the studies.

Preliminary Synthesis (Inclusion and exclusion criteria)
We included studies conducted from 1992 to June 2013 that reported the effects of health systems factors on hypertension control in SSA. The year 1992 was chosen as it was the year the first comprehensive evaluation of hypertension burden was conducted in the region [28]. To be included, and using the conceptual model as a guide, studies had to focus on key health system level inputs (See Fig 1) influencing hypertension management defined as: hypertension awareness, treatment, control, and medication adherence. Studies were included in the review if they met the following criteria; 1) distinctly identifiable study conducted in sub-Saharan Africa; 2) A direct measurement of any hypertension outcomes and aspects of health systems; 3) Studies conducted with a distinct population (patients, health care workers etc), populations on treatment, and /or with specific co-morbidities such as diabetes. All titles and abstracts were initially screened by two authors to determine eligibility. Disagreements were resolved by a third reviewer with expertise in hypertension management and the role of health systems. We excluded review articles, book chapters and dissertations. Also, we did not apply any language restrictions and the abstracts of articles conducted in other languages were initially scanned for relevance and translated to English if matching the inclusion criteria.

Data Extraction and Exploration of the Relationships between and within Studies
Using a standard template, the first author extracted the following information from available studies: (1) publication citation; (2) country where study was conducted and description of study participants; (3) study design; (4) description of hypertension outcome; (5) and aspects of health care system explored. Extracted data was reviewed by the remaining authors for disparities, erroneous and/or inconsistent data. Relationships between and within studies were explored further through thematic analysis to identify emerging themes relative to health systems building blocks and hypertension outcomes.

Quality Appraisal; Assessing the robustness of synthesis
The quality of the reviewed papers was assessed by two reviewers (JI and CE) using a checklist adapted from a previously described assessment criteria [29]. On the basis of the criteria, each reviewed paper received a quality grade of low, medium, or high, and any disagreement resulted in joint review of the article. The reviewers specifically appraised the data for the following:

Results
PRISMA guidelines were used to guide reporting of the literature reviewed and a flow diagram is shown in Fig 2 (See S1 PRISMA Checklist for more detail). Seven hundred and ninety-three articles were screened by title and abstract for inclusion. Out of 793 articles, the full text of 187 articles were obtained and assessed for eligibility. 25 studies met the eligibility criteria for this review. Full details of the included studies, including study design, population, setting, key findings and risk of bias assessment can be found in Table 2. 18 studies were quantitative in nature, four qualitative [30][31][32][33], and two were mixed-method study [34,35]. Of the 18 studies, one was a cluster-randomized trial [36]; 4 were cohort studies [37][38][39][40]; one of which was retrospective [39] alongside 2 other retrospective studies [41,42] and one mixed-method retrospective study [35]; 10 were cross-sectional studies. Majority of the studies were carried out in South Africa (n = 9) and in Nigeria (n = 8), 4 in Cameroon, while the rest in the Democratic Republic of Congo, Ethiopia, Ghana, and Tanzania. The narrative synthesis identified the Narrative Synthesis, Health Systems, Hypertension Sub-Saharan Africa  following themes (Table 3): 1) how physical resources within and outside the health system influence supportive mechanism for optimal hypertension control; 2) the role of human resources with enabling and/or inhibiting hypertension control goals; 3) the availability and/or use of intellectual resources; and 4) how health systems financing facilitate and/or compromise access to products necessary for optimal hypertension control.

Physical Resources
The findings identified how physical resources both within and outside the health care system influenced effective management of hypertension in region [4]. The physical resources operating within health systems include; unreliable drug supply, longer waiting times for appointment, the lack of functioning equipment, and poor referral and/or documentation of hypertension management. In Tanzania, Peck and colleagues [43] observed that first line drugs for hypertension were lacking and "even hospital outpatient departments sometimes did not have essential instruments such as sphygmomanometers." In the Agincourt sub-district of South Africa, Thorogood and colleagues [34] found that unreliable drug supply were among the barriers influencing the provision of effective treatment for hypertension. Clinic nurses interviewed noted that "they either had to deny treatment to patients or switch the treatment to another drug, actions that are likely to reduce adherence to medication." The nurses also reported that "they were unable to monitor blood pressure levels in those with hypertension due to lack of proper equipment in the clinics and/or non-functioning sphygmomanometers." In Nigeria, Mendis and colleagues (2004) observed that laboratory and other investigations to exclude secondary hypertension or to assess target organ damage were not available in majority of facilities. Physical resources operating outside the health system include; distance to clinic, particularly long travelling distance. in northwest Ethiopia, longer distance to hospital was found to be significantly associated with adherence to hypertension treatment recommendations with patients from long distant areas less likely to be adherent as compared to patients who resided closer to hospitals [44]. Finally, participants in a qualitative study in South Africa stated that travelling to the clinic where they received medication was extremely problematic. Coupled with the long travelling distance, absence of public transport and missing a day's wage, was a major impediment to attending clinic visits regularly [33].

Human Resources
The findings on human resources (trained health care workers) and optimal management of hypertension in SSA were mixed. For example, studies have found that task-shifting to nonphysician clinicians for integrated management of hypertension is feasible and effective with improving hypertension care in SSA. In a cohort study conducted by Kengne and colleagues [38] in rural and urban Cameroon, the authors implemented a nurse-led hypertension management protocol with 454 patients. Between baseline and follow-up, the authors observed that systolic blood pressure and diastolic blood pressure rates dropped by 11.7 mmHg (8.9-14.4) and 7.8mmHg (5.9-9.6) respectively. Similarly, in a study conducted by Ndou and colleagues [35], patients receiving home visits from community health care workers were more likely to achieve blood pressure control goals than those receiving usual care for from the clinic. On the other hand, physician inertia and/or failure to intensive antihypertensive medication are common in the region. In southwestern Nigeria, Ono and colleagues [39] observed that physicians increasingly "placed patients on monotherapy or "no drug treatment",. even in cases of uncontrolled systolic blood pressure." In the Gauteng Province of South Africa, Ndou and colleagues [35] observed that insufficient monitoring of patient outcomes by clinical staff hampered efforts to improve the management of hypertension [35]. Also in South Africa, among physicians interview by Parker and colleagues [45], observed that therapy was not routinely intensified for the majority of patients [45]. Similarly, in a cross-sectional study of blood pressure control in 451 hypertensive patients in South Africa, Rayner and Schoeman [46] found evidence of physician inertia with 60.6% of high-risk patients not at target and only 46.1%, 59.6% and 56.8% receiving advice about weight loss, exercise and diet, respectively. Moreover, the authors noted that in the patients not at target BP, there was no action plan in 22.9%, despite the fact that physician had self-identified these patients not to be at target and the majority of patients were either on one-or two-drug treatment [46].

Intellectual Resources
Three studies documented the availability and/or use of intellectual resources such as hypertension treatment guidelines. In South Africa, Daniel and colleagues [30], observed ambivalence towards the South African hypertension treatment guidelines noting "that physicians do not regularly consult the guidelines citing insufficient resources and time, overcrowded clinics, decreasing staff, and few opportunities for continuing medical education (CME)." Parker and colleagues [45] also observed that in South Africa, physicians were not conversant with the contents of the guidelines. In cases of adverse drug reactions, Yusuf and Balogun [41] observed that there was no 'institutional system in place to monitor, detect and document adverse drug reactions among patients on anti-hypertensive drug therapy.'

Social resources
No study reported on the impact of social capital on hypertension awareness, treatment and control in SSA.

Health System Financing
The findings on health systems financing which ranged from medication co-payments, costs of clinic visit to health insurance coverage were also mixed. A recent study among 3,023 community dwelling adults by Hendriks and colleagues [37] found that among patients in the community-based health insurance coverage program, systolic blood pressure decreased by 10.41mmHg (-13.28 to 7.54) and diastolic blood pressure by 4.27mmHg (-5.74-2.80). Despite these promising findings, health insurance coverage is often not available in many settings in SSA. Mendis and colleagues [47] noted that 'the fact that most patients have to pay out of their own pocket for consultation and medications either fully or in part, is likely to have a negative impact on long-term management of hypertension. In a survey of 250 rural patients with primary hypertension in South West Nigeria, Ilesanmi and colleagues [48] observed that about half of the patient were spending a tenth or more of their income on health care related expenses. Similarly, in a survey of 440 community residents in southwest Nigeria, participants stated that "medications were expensive compared to their income and they only bought the quantity of medication they could afford instead of the full prescription (for example, two weeks instead of four weeks)" [49]. High costs of medications have been shown to be associated with poor compliance to antihypertensive treatments in Ghana [32]. Buabeng and colleagues [32] found that in a sample of 128 patients, 93% did not comply with their medications, citing unaffordable drug prices as the main reason for non-compliance. However, one study found that even in cases where people are given access to free antihypertensive medication, hypertension was not controlled in about 38% of participants "suggesting that availability of free drug alone is not enough to improve adherence to antihypertensives [50]." A recent study conducted in Cameroon examined the association of medication costs with hypertension awareness and did not find one, although the confidence intervals were wide (odds ratio for hypertension awareness for high medication cost versus low medication cost 0.44, 95% CI 0.07-2.75) [51].

Governance and Service Delivery
No study examined health system arrangement and hypertension management relative to governance. But studies on physician inertia, such as the study conducted by Dennison and colleagues [52] in South Africa highlighted the influence of service delivery on hypertension management. For example, the authors suggested the following-"healthcare provider inertia sends a powerful message to patients, families, and other health care providers that hypertension control is not important." [52]

Discussion
The literature reviewed provides evidence on the health systems factors influencing hypertension awareness, treatment, control, and medication adherence in SSA. The findings showed that physical resources coupled with human and intellectual resources alongside health systems financing act in varying ways to influence optimal hypertension management in the region. Although barriers to optimal hypertension control exists at the level of the patients for example, in relation to knowledge about hypertension and its consequences, over and beyond these patient-level barriers [9,53], the findings of the literature reviewed suggest that systems level factors in the form of physical, human, intellectual resources and health systems financing enable and/or hinder equitable access to optimal hypertension care. Systems-level barriers such as lack of finances, high costs of drugs, longer waiting times at health systems, limited capacity for adequate diagnosing and prescribing, poor tracing of non-attending patients, and even poor provision of care by healthcare providers themselves also disproportionately limit the capacity of certain countries in SSA to manage and control increasing hypertension rates in the region.
The ability to provide optimal services by health care workers are influenced by increased patient load, overall increase in workload, and shortage of workers [54]. One key issue is the brain drain that has plagued SSA for over 20 years [55,56]. It has led to an acute and chronic shortage of health workers available to implement primary and secondary prevention of hypertension at available health care systems [54]. Although the literature reviewed provided information on the current state of knowledge regarding human resource that influence optimal management of hypertension, many gaps in the literature remain. For example, few studies examined the current workload of available health care workers or how their compensation rates may influence their ability to provide quality care for optimal hypertension. Moreover, the lack of available health care workers in most health care settings in SSA calls for new strategies to reduce brain drain. This knowledge base is necessary if reduction in hypertension related morbidity and mortality rate in SSA is to be achieved.
The same is obviously true for financial barriers in the form of high costs of medication for hypertension and overall health systems costs. For optimal control of hypertension to be achieved, considerable investments need to be made to improve access to equitable healthcare services without imposing excessive burden to individuals and households in the region. The availability of anti-hypertensive drugs is crucial for long-term treatment and control of hypertension. But studies also indicate that free antihypertensive medication is not sufficient by itself alone without the provision of other services such as the availability and use of accurate blood pressure monitoring devices or blood pressure treatment guidelines. Overall quality of care including acceptability (shorter waiting times), affordability (health insurance coverage or low clinic costs), accessibility (shorter distance to health care systems) also matters given the potential for low utilization of health care systems for adequate management of hypertension or non-compliance with antihypertensive drug therapy.
Taken together, delineating the health systems factors influencing optimal hypertension control in the region is a necessary first step towards the development and implementation of interventions aimed at mitigating the systems-level barriers to optimal hypertension control. It is also necessary for researchers to accurately capture the complexities inherent in resourcelimited health care systems in SSA and how this may have anticipated or unanticipated effects with hypertension awareness, treatment, control, and medication adherence in the region. While the reduction of hypertension in SSA is influenced by multiple factors operating at the health systems level, identifying key leverage points whether at the level of physical, human, intellectual or health systems financing level (instead of targeting only one level), is likely to produce larger and longer lasting effects.

Limitations of research
There are several limitations for the reviewed research. First, it is impossible to assess a direct link between health systems variables and the increasing prevalence of hypertension in the region because most of the available data are cross-sectional in nature. Longitudinal studies are recommended to understand the full impact of health systems on the management and control of hypertension in the region. Another important issue considered is that objectives of the available literature including the sampling methodology used were highly variable which in turn may raise concerns about the quality of the studies or the logic of combining the findings of all the studies reviewed. The use of narrative synthesis instead of a quantitative, meta-analytic review is crucial as it provided the opportunity to use words and text to summarize and explain findings from the literature reviewed thus providing evidence on the health systems factors influencing optimal hypertension in region. Furthermore, given that the findings were synthesized across countries with few countries represented in this review, the potential for regional differences within SSA may exist with health systems factors associated with optimal hypertension control varying significantly both within and between countries. However, this is a first attempt to discuss the context-specific complexities inherent in certain resource-limited health systems in SSA and how these might influence attempts to mitigate the increasing prevalence of hypertension in region. While a key challenge involves developing ways to understand and account for this complexity in a way that makes intervention practice effective, future research with key stakeholders that utilizes systems thinking strategies such as concept mapping for public health research or systems dynamic modelling, will enrich understanding and elucidate the multiple pathways through which health systems influence optimal management and control of hypertension in the region.

Conclusion
This narrative synthesis represents an effort to summarize existing data on the health systems factors that influence optimal hypertension control in SSA. The findings illustrate how physical, human, intellectual resources alongside health systems financing interact to influence optimal management and control of hypertension. While there were several methodological shortcomings identified, given the increasing prevalence of hypertension in SSA, the findings of this review underscore the need for conducting much-needed research as well as developing and implementing new interventions on a topic of vital public health significance to the region. Understanding health systems factors is also relevant for understanding the patient-level and/ or provider-level barriers that influence awareness, treatment, control, and medication adherence is the region. Furthermore, while there have been significant research effort to explore patient-level or provider-level factors, the success of new interventions for optimal hypertension control will also rely on an understanding of factors operating at the health systems level. Simply put, research or new interventions to reduce the increasing prevalence of hypertension in the region might be futile if the same vigor is not applied to addressing the health systems factors that influence the burden of hypertension in SSA. Moreover, knowledge of these factors have implications for the design and implementation of cost-effective interventions aimed at achieving optimal hypertension control in the region while facilitating the reduction of hypertension-related morbidity and mortality rates in SSA. This paper serves as a necessary first step towards open discussions on the complexities inherent in resource-limited heath systems in sub-Saharan Africa and how that might influence management and control of the increasing prevalence of hypertension in the region.