Prevalence of hypertension and its treatment among adults presenting to primary health clinics in rural Zambia: analysis of an observational database

Hypertension and its long term consequences constitute a growing burden in developing countries, with as much as 75 % of all affected patients now living in low- and middle- income countries (LMIC) [1‐5]. Defined by the World Health Organization as a blood pressure measurement ≥140/90 mmHg, hypertension is a major known risk factor for chronic conditions such as cardiovascular disease, hemorrhagic stroke, and renal failure [1, 6‐9]. According to the 2012 WHO Global Health Observatory, 46 % of deaths in LMIC are attributable to cardiovascular disease (CVD) [10]. Notably, CVD deaths occur in sub-Saharan Africa at a younger mean age (64.9 years) than anywhere else in the world [7].

In Zambia, non-communicable diseases (NCDs) account for up to a quarter of all deaths, with about half of these due to CVDs [11]. As life expectancy increases, this burden is expected to rise as well. There is a paucity of epidemiologic data on hypertension prevalence and treatment in Zambia [12, 13]. Like the vast majority of studies in sub-Saharan Africa, past studies in Zambia have used cross-sectional community survey designs with a single blood pressure (BP) measurement to define hypertension [12‐15]. To date, there are no studies examining the profile of hypertension patients who actually present to primary health clinics. Similarly, while control of hypertension with appropriate medication can greatly reduce complications like stroke and other cardiovascular events [16], there are very few studies that have examined and reported treatment practices.

To address these gaps, we sought to determine the age standardized prevalence and existing treatment practices of hypertension in 46 Zambian rural primary health care centers participating in the Better Health Outcomes through Mentoring and Assessment (BHOMA) intervention [17].

The BHOMA project is a five-year, randomized stepped-wedge trial of improved clinical service delivery underway in 46 rural clinics in Chongwe, Kafue, and Luangwa Districts. The public sector lacks robust healthcare infrastructure and suffers from a severe human resource shortage, with only 0.7 physicians and 7.8 nurses per 10,000 people [18]. A number of our selected rural clinics are staffed by one nurse or clinical officer, and frequently suffer from equipment shortage. BHOMA has five principal objectives: 1) creating standardized protocols and forms for common primary care visits, 2) ensuring clinics are properly equipped with essential diagnostic and management tools, 3) on site electronic medical records (EMR), 4) improving key indicators with ongoing mentoring, and 5) increasing community engagement with community health workers and traditional birth attendants [17]. This study is an analysis of the observational EMR database.

Between January 2011 and December 2014, a total of 1,021,530 visits were documented at 46 participating BHOMA health facilities. Of these, 318,384 (31.2 %) were primary care visits by adults ≥25 years. After excluding repeat visits and first visits without a BP measurement, a total of 116,130 first visits by adults were left (median number of visits: 1; interquartile range: 1,2). This comprised our analysis cohort (Fig. 1). A total of 48,843 (42.1 %) adults in the cohort made more than one visit.

Fifty seven percent of patients were female, with an overall mean age of 40 years dominated by younger participants under 45 (69.7 %). Most patients (71.6 %) had a body mass index (BMI) <25 kg/m² (Table 1). Our adult study population was significantly older (mean 40 years, median 36 years) compared to the overall population (mean 20 years, median 18 years) that presented to BHOMA facilities in the study timeframe.

In our rural primary care setting, similar to other studies in the region, we observed a high prevalence of hypertension among adults seeking care. Increasing age and BMI were associated with increased OR of hypertension. Male gender was associated with increased OR only in the youngest age band, with a lower OR in older age bands compared to females. Treatment coverage was low with only 18 % of patients with hypertension prescribed medication. Individuals with uncontrolled hypertension were more likely to be female, had higher BMIs, and interestingly more likely to be prescribed nifedipine compared to those with controlled hypertension. However, in multivariable regression, only the highest age category ≥45 years and BMI category ≥30 kg/m² were associated with a statistically significant decreased OR of hypertension control.

This study had several limitations. First, external validity is a concern since our study sampled individuals who sought healthcare at primary care facilities, rather than the population as a whole. Consequently, certain results such as the age standardized prevalence of hypertension may be lower than a community age standardized prevalence. Second, only data from the first visit was used from the longitudinal dataset to arrive at prevalence. Similar to other cross sectional studies, BP measurements from one visit may be spuriously elevated, leading to an overestimation. This trend was evident when we considered the median BP over all visits; in that sensitivity analysis, the overall age-standardized prevalence of hypertension was slightly lower compared to first-visit measurements only. Third, the specific indication for a prescribed drug was not known—for example, we could not confirm from the data whether a loop diuretic was prescribed for heart failure or hypertension. Fourth, key independent variables for inclusion in the regression models (smoking status, family history of hypertension, diet, exercise) were not collected as part of the study. These were balanced against the study’s strengths. Data collected at health facilities represented how closely health centers put theoretical guidelines into practice, and are more relevant for informing interventions integrated into existing infrastructure. In addition, calculation of age standardized prevalence allowed comparison to other prevalences standardized to the WHO World Population.

Within Zambia, our crude prevalence of 23.1 % was lower than the crude prevalence found in previous studies in Lusaka urban district of 34.8 % (38.0 % male, 33.3 % female), but closer to the proportions found in rural Kasama district of 30.3 % (31.3 % male, 29.5 % female), and rural Kaoma district of 25.8 % (27.5 % male, 24.6 % female) [12, 13]. The difference between Lusaka urban district and our data was likely due to our rural sample, which presumably has fewer dietary and lifestyle risk factors for developing high blood pressure than urban samples [1, 15]. The prevalence of hypertension in sub-Saharan Africa has ranged from 10–55 % [14, 15]. Our crude prevalence was slightly lower than the urban areas in neighboring countries Zimbabwe (33.3 %) and South Africa (25.0 %), but higher than rural areas in South Africa (10.5 %) [1, 21]. Differences in sampling, or underlying population age and gender distributions, may have contributed to the wide range of reported crude prevalences. Our age standardized prevalence of 28.0 % was similar to the age standardized levels reported by the Millenium Villages in rural Malawi, Rwanda, and Tanzania (22.8 % hypertension) [22].

Similar to other studies in rural sub-Saharan Africa, our proportions for hypertension by measurement were higher in females than males—especially in older age bands. A meta-analysis showed 8 out of 13 total rural African communities found a higher hypertension prevalence in females compared to males [14]. The location of the effect in older age bands, compared to the younger groups, is consistent with research in other world regions [23, 24]. It is thought that older, postmenopausal women lose the protective role of ovarian estrogen on vasodilation and decreased sympathetic activity, thereby resulting in an elevated BP [25]. Female predominance was also more apparent in the hypertension by diagnosis, which is consistent with prior research demonstrating females were more likely to have been detected, to be on treatment, and have hypertension controlled compared to males [14]. Females have more touch points with the healthcare system through antenatal care visits and have better health seeking behavior, meaning they may present more frequently to health centers and be more likely than males to be detected [14, 26, 27]. Indeed, in BHOMA the majority of patients were females (57 % of study cohort vs 51 % of population) [28].

Another major finding in our study was the under- and improper treatment of hypertension. Zambian national guidelines, consistent with the Eighth Joint National Committee criteria, state that after unsuccessful lifestyle modification, antihypertensive treatment should start with one agent (hydrochlorothiazide-amiloride or a calcium channel blocker), with increasing doses or combinations if BP remains elevated [29, 30]. Almost all classes of antihypertensive drugs are on the essential medicines list and provided free in Zambian government clinics, including thiazides, calcium channel blockers, and ACE inhibitors. A major contributor to under-treatment seemed to be under-diagnosis; another contributing factor may have been healthcare worker discomfort with chronic diseases, as 50.7 % of Zambian government health workers feel a lack of expertise in managing hypertension [31]. While under-treatment certainly contributes to persistent hypertension, even the vast majority of treated patients had uncontrolled hypertension. Possible reasons include lack of awareness in the general population about dietary salt contributions to elevated BP, poor patient adherence or clinic attendance for refills, and improper treatment choices by clinic staff. Furosemide was commonly used (in 8 % of hypertensive patients), which is only recommended in the presence of renal or cardiac failure [32]. Given the extremely low numbers of documented cardiac failure in our sample, it is unlikely that patients were evaluated for these conditions at the primary care level, and thus unlikely to be on furosemide for correct indications. Improper treatment is likely tied to many factors, including medication stock outs and lack of training in best practices.

Our study showed the high prevalence of hypertension in Zambian primary care clinics and the need for early intervention. Given limited resources, interventions should focus on population-level and individual-level public health approaches in parallel. Passing national legislation on hypertension risk factors (tobacco, alcohol, poor diet) may have a cost effective yet considerable public health impact. Modeling in LMIC shows that a 33 % tax on tobacco could result in only US$ 3–42 spent per DALY averted from cardiovascular disease due to decreased smoking [33]. At the individual level, emphasis must be placed on identifying those at highest risk and ensuring proper treatment. The WHO and International Society of Hypertension have developed low resource wall charts of the Framingham cardiovascular risk calculator, requiring only gender, SBP, smoking status, and diabetes mellitus type II status [34]. Treating only patients with a total cardiovascular risk of >15 % in primary prevention can be cost effective at US 790-930 per DALY averted, and is cost saving for secondary prevention [33]. Our study also demonstrated individuals were being treated with the ineffective drug furosemide, suggesting the need for broader health systems strengthening in prescribing practices and drug stock outs. There remain many unanswered questions about hypertension in Zambia, including reasons for the poor screening and treatment in the BHOMA clinics (despite its objective to improve quality of care through structured protocols), treatment practices at higher hospital levels of care, and evaluation of NCD training and practices. Future studies will need to explore reasons for hypertension under-treatment to develop sustainable interventions for primary and secondary prevention of cardiovascular disease.

In summary, in our rural Zambian primary care setting, the age standardized prevalence of hypertension was high at more than one-quarter of adults aged ≥25 years. Consistent with other studies, females were noted to have a higher prevalence than males especially at older ages, likely due to greater engagement with the health system and menopausal status. Treatment coverage was low, with evidence of improper prescriptions for those who were treated. To prevent progression of hypertension to cardiovascular mortality, future interventions should implement dual population preventive care and high risk individual strategies.