Self-reported hypertension as a predictor of chronic health conditions among older adults in Ghana: analysis of the WHO Study on global Ageing and adult health (SAGE) Wave 2

Introduction Hypertension has been identified as a significant predictor of many chronic health conditions. Body Mass Index (BMI) and Quality of Life (QoL) are key determinants of hypertension especially among elderly populations. In this study, we examined the effect of self-reported hypertension (SRH) on chronic health conditions and quality of life among older adults in Ghana. Methods The WHO Study on Global Ageing and Adult Health Wave 2 data for Ghana, collected from 2014 to 2015 was applied in this study. Data for older adults aged 50 years and above were analyzed. Weighted descriptive and inferential analyses were performed using Stata 14. We predicted any potential associations between SRH and chronic health conditions using a corrected chi-square and Coarsened Exact Matching with adjusted odds ratios. Results The prevalence of SRH among older adults in Ghana was 15.8%. This was significantly associated with sex, marital status, religion, place of residence, working status, location/region, health status BMI, and QoL. In all, older adults with poor health status, obese state and high QoL had 3.15, 2.17 and 2.76 odds of SRH respectively [AOR(95%CI)p-value=3.15(1.65-6.02)0.001, 2.17(1.31-3.59)0.003 and 2.76(1.04-7.31)0.041)]. In addition, older adults with SRH were at increased risk of reporting chronic conditions such as stroke, angina, diabetes and cataract. Conclusion Overall, a key observation from this analysis is that SRH (and not only clinically diagnosed hypertension) is significantly associated with co-morbidities. In Ghana, older adults with SRH have increased risk of co-morbidities including diabetes, stroke, angina, and cataract. Interventions to improve the awareness and early detection of hypertension at the population level is key. Controlling hypertension at the population level will reduce prevalence of chronic conditions and increased protection.


Introduction
Hypertension is a major risk factor for cardiovascular diseases worldwide [1]. The 2017 Global Burden of Disease (GBD) indicated that hypertension was an important public-health challenge worldwide and was estimated to affect 1.56 billion individuals by 2025 with an increased global prevalence of 60% [1,2]. Hypertension is a chronic multifactorial disease whose detection is often delayed due to its slow and silent progression [2]. However, this asymptomatic chronic disease in most people requires optimal control usually through disciplined adherence to prescribed medications. Without optimal control, there is increased risk of developing cardiovascular, cerebrovascular and renal complications [3].
In sub-Saharan Africa (SSA), the prevalence of hypertension and prehypertension is high, with an estimated prevalence rate of 25.9%.
This, however, differs by population group [4]. The 2014 Ghana Demographic Health Survey (GDHS) reported a hypertension prevalence rate of 13% among adults [5] in the reproductive age, while the WHO Study on Global AGEing and Adult Health (SAGE), Wave 1 in Ghana reported SRH prevalence rate of 14.2% among older adults 50 years and above [6]. Age is universally associated with hypertension, and in Africa, the greatest burden of hypertension is borne by persons 50 years and above [7]. Increasing age, is a major cause of reduction in the quality of life (QoL) among older adults. [8].
The QoL which is an individual´s perception of happiness and satisfaction with life as well as position in life and value systems in relation to expectations, values and concerns with physical health may be severely compromised [9]. Low QoL affects the psychological well being and socio-demographic functions [10]. Invariably, QoL is an important predictor of treatment outcomes among older adults with hypertension and QoL among older adults with hypertension has been shown to be worse than those without hypertension [10,11].
However, hypertension among older adults may not be detected early or may not be adequately managed [12]. The consequences of this on the health and social well being of these older persons may be dire.
Imperatively, measures needed to be taken for early diagnosis of hypertension and to prevent its complications is essential at the population level, considering that the prevalence of hypertension is estimated to increase [2]. Globally, there is substantial connection between hypertension and other chronic health conditions which reflect in their etiology and disease mechanisms. Hypertension is common among patients with diabetes [13] and stroke [14] . It has been observed in previous studies that 51% of stroke deaths are attributable to hypertension [15]. Hypertension is associated with other chronic conditions that affect the QoL of older persons including eye conditions such as cataract. This impairs the daily functioning of the older adults and affects their quality of life [16]. This analysis determined the prevalence of SRH, predictors of hypertension and quantified the effect of SRH on stroke, angina, diabetes and cataract among older adults in Ghana.  [6,20]. Detailed study design and procedure for data collection adopted in SAGE wave 2 is well described by Kowal et al.

Methods
(2012) [21]. A total of 4,735 persons, comprising older adults (50+ years) and persons in their reproductive age (18-49 years) were interviewed. In line with the aim of the study, data for persons below 50 years and missing responses for SRH were excluded. A final sample of 2,335 adults aged 50 years and above were included in this study.
Dependent variables: self-reported hypertension (SRH) was considered the primary outcome variable.. Co-morbidities including stroke, angina, diabetes and cataract were also considered as outcome variables. The SRH was defined 'as ever been told by a doctor or health care professional that you have high blood pressure (hypertension)´. Independent variable: from the original dataset, sex, marital status, religion, educational attainment, residence, region of residence and employment status were extracted for the sociodemographic variables. BMI was estimated from a measured weight (kg) and height (metres) and categorized into <18.5, 18.5-24.9, 25-29.9 and 30 and above as underweight, normal, overweight and obesity respectively. Coarsened Exact Matching (CEM) method of analysis was carried to improve and reduce imbalances in estimating causal effects. This was conducted to control for some or all of the potentially confounding influence of pretreatment control variables by reducing imbalances between the treated and control groups. This Monotonic Imbalance Bounding (MIB) matching method which balances between the treated (SRH) and control groups (non-SRH), was employed to adjustthe influence of the imbalance of one variable on the maximum imbalance of any other [22]. In this analysis, SRH predictors identified included; Age, Place of residence, Region, Health status, BMI and QoL.
These were modeled to reduce the imbalances between them (Table   1).
Overall 76% of imbalance existed among the independent variables before matching. However, after CEM matching, imbalances reduced to almost 0% (5.63E-16) ( Table 1). After preprocessing the data with CEM, sensitivity analysis was applied involving Logistics regression, weighted with CEM estimations to estimate the risk of SRH on stroke, diabetes, angina and cataract adjusting for non-predictors of SRH which includes; sex, marital status, religion and working status. The concept and process involved in conducting this analysis is demonstrated in Figure 1; that older adults with SRH can either be controlled with medication or uncontrolled without medication. The risk of stroke, angina, diabetes and cataract can be either low or high among older adults with SRH in relation to management of the condition, whilst the demographic predictors can also directly affect the health condition.

Results
Data on a total of 2,335 older adults 50 years and above were analyzed ( Table 1). The study found a prevalence of 15.8% for SRH among the older adults in Ghana. Sex differential in SRH existed and was relatively higher among females (20.3% vs. 11.6%). Those in the age group 60-69 years had the highest prevalence of SRH (18.5%).
Older adults who reported marital status as separated, had the highest prevalence of SRH (22.5%). In addition, rural-urban disparities existed in the prevalence of SRH and was relatively higher among urban residents (23.1% vs. 8.1%), while unemployed older adults had relatively higher SRH, than those who were working (19.8% vs. 13.2%) In terms of geographical location, older adults living in the Volta region had the highest risk of SRH (51%), compared with those living in the other nine administrative regions in Ghana.
Unsurprisingly, poor perception of health status was associated with a higher SRH among the older adults (28.7%) and SRH was highest among obese individuals (34.2%) as compared with other BMI categories ( However, this finding contradicts to an observation that. single individuals are better off in blood pressure profile than unhappy married individuals [33]. Intuitively, emotions and stressors in such a union may increase the risk of hypertension. It has been established that religion improves QoL and lowers risk for developing hypertension [34]. Aligned with this, religion was found to be significantly associated with SRH and that older adults without any form of religious affiliation reported a higher prevalence of SRH (18.3%). Religion, belief systems and health have been demonstrated to be related in all population groups and is statistically correlated to the prevalence of hypertension [35,36].
Generally, there are inconsistencies in rural-urban disparities in the prevalence of SRH. In our analysis, urban dwellers had relatively higher SRH while rural-dwelling adults were at lower risk and had a decreased predictive probability of SRH. This finding is consistent with the findings of several publications [37][38][39][40]. At the same time, it contradicts to findings of other published work where the burden of hypertension among the older adult is significantly higher in rural areas [12,[41][42][43][44]. In Ghana, the relative higher prevalence of SRH in urban areas is not surprising. Rural dwellers are more active, are engaged in more physically tasking occupations and have lower prevalence of obesity and inequities in access to health providers [19].
Conversely, urban dwellers may have increased access and exposure to information (traditional and social media) and better awareness of hypertension in the urban areas. This is buttressed by the finding that the most urbanized region, Greater Accra, had the highest prevalence of SRH in Ghana.
In this analysis, adults who were not working reported a greater prevalence of SRH which was similar to a study conducted in Japan where job strain was significantly related to hypertension [45].

Conclusion
Overall, a key observation from this analysis is that SRH (and not only clinically diagnosed hypertension) is significantly associated with comorbidities. In Ghana, older adults with SRH have increased risk of co-morbidities including diabetes, stroke, angina, and cataract.
Interventions to improve the awareness and early detection of hypertension at the population level is key. Controlling hypertension at the population level will garner reduction in prevalence of chronic conditions and increased protection.
What is known about this topic  14.2% self-reported hypertension exist among older adults 50 years and above;  The 2014 Ghana Demographic Health Survey reported a hypertension prevalence rate of 13% among adults;  There is substantial connection between hypertension and other chronic health condition.

What this study adds
 The prevalence of self-reported hypertension among older adults in Ghana was 15.8%;  Self-reported hypertension is associated with sex, marital status, religion, place of residence, working status, location/region, health status, BMI, and quality of life;  Older adults with self-reported hypertension were at increased risk of reporting chronic conditions such as stroke, angina, diabetes and cataract.

Competing interests
The authors declare no competing interests.

Authors' contributions
JT conceptualized the study and sought approval for access to the  Table 1: CEM weighted balance report before and after matching